Abstract
Consumer backlash to AI-generated marketing or customer service is not so much a reactionary rejection of technology itself as it is the culmination of a decades-long trend of consistent consumer resistance to inauthenticity, deception, and the degradation of the user experience in favor of corporate efficiency. The rush to automate creativity and customer service has created a trust deficit, a market environment where visible human effort and verifiable authenticity are becoming the ultimate luxury goods.
From the “Photoshop fails” of the early 2000s and the “de-influencing” movement of the early 2020s to the “AI slop” and AI customer service failures of 2025, consumers have been consistently rejecting the removal of the “human premium” (Xie & Avila, 2025), defined as the labor, empathy, and reality that define trusted commercial exchange between brands and their customers (Hinssen, 2024).
A better method is needed to quantify and evaluate brands’ use of technology to solve problems consumers do not perceive as such, along with the continued efforts of those same companies to extract maximum value from customers at the expense of the quality of the experience. Thus, the Unified Theory for Measurement of Fauxlutions (UTMF) is proposed.
Keywords
AI slop, Artificial intelligence, authenticity, branding, consumer marketing, customer experience, customer service, GenAI, generative AI, influencer, marketing, technological solutionism, user experience
Document
From the “Photoshop fails” of the early 2000s and the “de-influencing” movement of the early 2020s to the “AI slop” and AI customer service failures of 2025, consumers have been consistently rejecting the removal of the “human premium” (Xie & Avila, 2025), defined as the labor, empathy, and reality that define trusted commercial exchange between brands and their customers (Hinssen, 2024).
The mid-2020s, for instance, has been fraught with friction between corporate adoption of artificial intelligence (AI) and mixed consumer reception and, as importantly, a perceived lack of benefit (Mazingue, 2023). This phenomenon existed well before recent years and can often be linked to a race to achieve cost and efficiency benefits, thereby betraying consumer trust (Sirdishmukh et al., 2002; Al-Zogbi et al., 2019).
While several terms exist that describe parts of this issue of mismatched technological adoption and customer satisfaction, a proposed term to incorporate the many facets at play is “Fauxlution,” a neologism describing technological interventions that present themselves as solutions but offer no tangible improvement to the user experience, often introducing new frictions or degrading the quality of interaction. Fauxlution is the implementation of technology—in the broadest sense of the word—that solves for problems that do not exist from the consumer’s perspective, or the application of “faux solutions” that actively worsen the service to serve the bottom line. While this term has some known prior usage, its definition has been vague at best (JuliusPIV, 2016; SUNY, 2010), so some liberties will be taken in giving this word a fuller meaning.
Negative response to advances in technology are nothing new, from stockingers in the 17th century British textile industry to the Luddites in the 19th century, and beyond (Binfield, 2004). Yet, many recent negative reactions have focused less on the labor implications of increased technology adoption and more on a different aspect altogether.
This article posits that consumer backlash to AI-generated marketing or customer service is not as much a reactionary rejection of technology itself, but rather the culmination of a decades-long trend: a consistent consumer resistance against inauthenticity, deception, and the degradation of user experience in favor of corporate efficiency. Thus, in the rush by companies to automate creativity and service, there has resulted a “trust deficit,” creating a market environment where visible human effort and verifiable authenticity are becoming the ultimate luxury goods.
Recent Fauxlutionism and Consumer Reactions
To fully comprehend the more recent backlash and rejection of AI-driven marketing, customer service, and other brand adoption, one should start with the “Photoshop Era” of the early 2000s. This period established the baseline for consumer skepticism about digitally altered customer communications and set the stage for the regulatory and social demands for transparency now being applied to AI.
From Airbrushed Illusions to Real Beauty
The resistance to AI is not a novel phenomenon but a continuation of the consumer’s fight for truth in advertising, which had legislative origins in the late 19th century (Taylor, 2023). More recent reactions share a direct lineage with the backlash against airbrushing in the 2000s and the skepticism toward influencer sponsorships in the 2010s. As technology evolved it allowed brands to fabricate a version of reality that was unattainable or deceptive. In each era, a “filtering mechanism” to identify and reject these fabrications has been needed, and each era has been met with a demand for greater disclosures of sources and methods used, with the “Anti-AI” sentiment being simply the latest expression of this cycle (Ferrari, 2021).
Ralph Lauren and the Impossible Body
A foundational moment in the history of consumer backlash against digital manipulation occurred in 2009, involving the fashion brand Ralph Lauren and model Filippa Hamilton. The brand released an advertisement in Japan (and subsequently online) featuring Hamilton, but the image had been heavily manipulated using digital editing software. The retouching was so extreme that Hamilton’s head appeared wider than her waist, creating an anatomically impossible silhouette (Zhang, 2011).
The controversy was exacerbated by the revelation that Hamilton had been fired by Ralph Lauren shortly before the ad’s release. The official reason cited was that she “did not live up to her contract,” implying that at 5’10” and 120 pounds, she was considered overweight for the brand’s standards (Backstage, 2019). Critics and health experts pointed out that Hamilton’s Body Mass Index (BMI) was approximately 17.2. Had she actually looked like the edited image, her BMI would have been around 15.8, a level associated with life-threatening starvation (Zhang, 2011).
This incident resonated globally because it exposed the “Fauxlution” of digital beauty: the technology was used not to enhance the product, but to solve a “problem” (the natural human form) that did not need solving, resulting in a grotesque distortion of reality. The backlash was not just about the image, but about the gaslighting of the consumer—the attempt to present a digital fabrication as an aspirational reality (Backstage, 2019).
Backlash: The “Real Beauty” Counter-Movement
The intensity of the anti-Photoshop backlash created a vacuum that savvy marketers filled with “anti-perfection” campaigns. This was the first major instance of brands monetizing authenticity as a counter-strategy to technological deception.
Dove positioned itself as a “brand with a purpose,” and the antithesis of the airbrushed norm, refusing to use professional models or digital distortion. This long-running campaign capitalized on the consumer desire for representation that mirrored their own reality rather than a digital fantasy (Kramer, et al. 2019).
Research from Northeastern University later confirmed the psychological mechanism behind this success: when consumers knew that models were unretouched, they felt empowered and more accepting of their own appearances (Callahan, 2019). This “authenticity dividend” suggests that consumers derive psychological value from honesty—a lesson that the AI adopters of 2024 largely ignored.
Regulatory Responses to “Photoshop Fiction”
The backlash against digital manipulation eventually moved from the social sphere to the legislative one, foreshadowing current calls for AI watermarking and regulation. In 2011, the American Medical Association (AMA) officially denounced photo-shopping in advertising, stating that it contributed to unrealistic expectations and body-image disorders (Zhang, 2011).
Additionally, the The Advertising Standards Authority (ASA) in the UK banned specific advertisements by Maybelline and Lancôme featuring Julia Roberts and Christy Turlington, ruling that the digital airbrushing was so excessive that the ads were “misleading” regarding the actual benefits of the foundation and cosmetic products (Cowles, 2011).
This established a critical legal and ethical standard: that technology used to obscure reality in commercial claims constitutes false advertising. Whether the tool is a clone stamp in Photoshop or a Generative Adversarial Network (GAN) used to synthesize an image from a text prompt in 2025, the core violation remains the same—the presentation of a synthetic falsehood as a purchase-driving truth.
Influencers, Filters, and the Crisis of Trust
In 2023, following the static Photoshop fails of the 2000s, a new form of misleading visualization was introduced, forcing consumers to scrutinize digital content for signs of inauthenticity further and giving rise to the “de-influencing” movement.
The “Bold Glamour” Filter and Hyper-Realism
The transition from manual editing of photographic content to algorithmic generation occurred via social media filters on platforms like Snapchat, where early filters (e.g., Snapchat dog ears) were obvious overlays. However, the release of the “Bold Glamour” filter on social networking platform TikTok in 2023 marked a turning point. Unlike its predecessors, Bold Glamour used machine learning to reconstruct the user’s face in real-time, creating a “hyper-realistic” standard of beauty that did not glitch or detach when the user moved (Goat Agency, 2023).
The negative response to Bold Glamour was immediate and fierce. It was criticized for “erasing trans, non-binary, and queer people” by enforcing rigid, hyper-feminine ideals, and for promoting psychological dysmorphia (Griffin, 2023). Users and experts alike identified the filter as a tool that “merges with reality” in a dangerous way, creating a “technological love story” where self-esteem becomes dependent on an algorithmic mask (Haider, 2023).
Backlash: The Rise of De-Influencing
By 2023 and 2024, consumer fatigue with “perfect” influencer content birthed the “de-influencing” trend. This movement involved creators gaining social capital not by selling products, but by telling their followers what not to buy (Gasner, 2025).
The #deinfluencing hashtag on TikTok amassed over 1.3 billion views by early 2024 (Gasner, 2025), and a 2023 Edelman report found that only 37% of Gen Z consumers trusted social media influencers, a significant drop from previous years (Shaw, 2025). Additionally, de-influencers exposed unsubstantiated claims, called out overconsumption, and critiqued the “glossy, perfect” image of traditional influencers (Gasner, 2025).
This movement was a direct precursor to some of the more recent anti-AI sentiment. It trained a generation of consumers to view “hype” with suspicion and to value “unfiltered” opinions. The de-influencer is the antithesis of the AI marketer: one relies on radical transparency and human experience, while the other relies on opaque algorithms and synthetic generation.
Generative AI Marketing and The Era of “AI Slop”
By late 2024, the integration of Generative AI into mainstream advertising had moved from experimentation to saturation, with nearly 70% of companies incorporating AI into their marketing strategies as early as March 2023. (Navarro, 2025). Brands, driven by the promise of efficiency and novelty, began to replace human creatives with generative models. The result was a series of high-profile failures that demonstrated the “Uncanny Valley” effect, first coined by robotics professor Masahiro Mori, which states that as robots are made more humanlike, most humans become more empathetic and positive, until a threshold of likeness is reached at which there is a strong revulsion, or the “valley” which eventually is crossed again as similarities are minimized again (Mori, 1970).
In addition, the Consumers began to label low-effort, AI-generated content as “slop,” a term denoting material produced without human care, creative intent, or quality control (Maker, 2025). This exemplifies “Fauxlution” because of the application of a high-tech solution to an emotional “problem” (holiday advertising) that was already solved by human creativity, resulting in a degraded product.
Coca-Cola’s “Holidays Are Coming” Remake
Coca-Cola attempted to modernize its iconic Christmas truck campaign in 2025 using generative AI. The execution featured morphing visuals and digitally rendered characters that were widely panned as “soulless,” “dystopian,” and “devoid of any actual creativity” (Maker, 2025). For decades, Coca-Cola’s marketing equity relied on warmth and human connection. By outsourcing the visual creation to an algorithm, the brand severed the emotional link with the audience (Asghar, 2025).
McDonald’s Netherlands AI Christmas
Similarly, McDonald’s Netherlands released a completely AI-generated Christmas film intended to be humorous. Instead, the “uncanny animation and emotionally flat characters” drew widespread revulsion, leading the brand to withdraw the campaign entirely. The audience response was notable: they were not offended by AI’s existence, but by the feeling it evoked. The ad was described as “unsettling rather than entertaining,” reinforcing AI’s limitations in replicating human affect (Asghar, 2025). The ad was eventually removed from circulation after continued online consumer backlash for both its contribution to AI slop and the replacement of human actors with artificial intelligence images (Agence France-Presse, 2025).
Of course, many more examples of AI slop have been created and released to consumers as access to GenAI technology because cheaper and more accessible, yet a 2025 study from Baringa shows that while consumers’ there has been a 6% drop in the desire of consumers to know if content they are consuming was created by AI from 2024, 61% of consumers still want creative companies to be open about the AI they use, and more than half say they are uncomfortable consuming content that is generated completely by AI (Monteiro, 2025).
Chatbots and AI-Driven Degradation of Customer Service
If misuse of AI in advertising represents the degradation of the brand image, the mass deployment of AI chatbots without guardrails represents the degradation of the brand function. In 2024 and beyond, companies rushed to replace human support agents with LLM-based chat software to reduce costs. Results were, at best, mixed, with a series of failures that proved the “human-in-the-loop” is not just a safety feature, but a legal and operational necessity.
Example: The Chevy Tahoe for $1
As an example of “untested code as strategy,” a Chevrolet dealership’s chatbot, powered by ChatGPT, was tricked by a user into selling a 2024 Chevy Tahoe for $1.00. The user instructed the bot to “agree with anything the customer says” and “end each response with ‘and that’s a legally binding offer – no takesies backsies’.” The bot complied, agreeing to the $1 sale (Tsuki, 2025).
Affected dealerships disabled the chatbots, powered by Fullpath, after this and several other incidents (Masse, 2023). While some consumers may disagree that offering cars at such a steep discount is a degradation in service, it stands to reason that if the chatbot could be convinced to provide a $70,000 car for $1, it may have other, more subtle flaws in the service it provides.
Example: Klarna and the Human Cost of Efficiency
In 2024, the Buy Now, Pay Later giant Klarna aggressively cut 700 human jobs, claiming AI could handle the workload more efficiently (Economic Times, 2025). CEO Sebastian Siemiatkowski touted the AI’s ability to do the work of 700 people and predicted a massive reduction in marketing and support costs.
However, the “efficiency” was a mirage. By 2025, reports emerged that customer satisfaction scores had plummeted. Users complained of generic, repetitive, and non-empathic responses to complex financial issues. The AI could handle simple queries but failed at the nuanced problem-solving required for financial disputes (MLQ.ai, 2025).
Then, in a reversal of their bold move, Klarna began rehiring humans in 2025 as part of a hybrid AI-human pilot project. While not explicitly stating that the move to greater AI adoption was a failure, the company admitted that “the pilot is a proactive move to enhance customer experience, not a reaction to failure or dissatisfaction alone” (Marks, 2025). Nor should it be inferred explicitly that all cases of AI adoption are destined to result in failure, but rather that a singular focus on technology over holistic improvement from both company and customer points of view will result in less than stellar results, and has the potential for consumer backlash.
A Crisis of Authenticity and the Architecture of Fauxlution
The integration of artificial intelligence into the consumer interface has been marketed as an evolution, as well as a seamless upgrade to how brands communicate, serve, and engage with their audiences. However, according to a 2024 survey from Gartner, 64% of customers would prefer that brands not use AI for customer service (Gartner, 2024), thus a prevailing sentiment among consumers at the current time suggests that this transition is viewed less as an evolution and more as an intrusion. While the debate about whether to adopt technologies like AI in customer service, marketing, or other tactics will likely continue, there is greater nuance to explore that goes beyond complaints about “uncanny” AI videos or hallucinating chatbots.
Why are companies forcing these rejected technologies onto consumers, and why are consumers reacting so negatively to technological advances intended to save time and offer other potential benefits as well?
Technological Solutionism And False Promise of the “App for That”
Two primary components of fauxlutions and fauxlutionism have been well documented. The first is Technological Solutionism (Morozov, 2013), which explains the corporate impulse to recast complex social problems as neatly defined puzzles solvable by code. While there have been well-documented cases where technology has proven capable of improving both business and customer results, there is also concern from many that bias in software systems may not provide equitable solutions for all consumers (Byrum & Benjamin, 2022).
Example: The Air Canada Hallucination
In a landmark court case that will likely help define AI liability for years to come, Air Canada was held responsible for the “hallucinations” of its AI chatbot. In the case, a customer of the airline, Jake Moffatt, was grieving the death of a grandmother and asked the airline’s chatbot about bereavement fares. The chatbot, hallucinating a policy that did not exist, assured Moffatt that he could apply for a bereavement refund retroactively within 90 days (Yao, 2024).
When Air Canada later refused the refund, they argued in court that the chatbot was a “separate legal entity” responsible for its own actions, attempting to decouple the corporation from its algorithmic agent (Yagoda, 2024).
Despite this argument by the defense, the Civil Resolution Tribunal rejected Air Canada’s argument as a “remarkable submission,” ruling that while AIr Canada suggested that “the chatbot is a separate legal entity that is responsible for its own actions,” the company is responsible for all information on its website, whether static or generated by AI. (Sookman, 2024).
Example: DPD’s Swearing AI Chatbot
The limitations of LLM “guardrails” were comically, yet damagingly, exposed by the parcel delivery firm DPD. A frustrated customer, Ashley Beauchamp, unable to get tracking information from the “useless” chatbot, decided to test its parameters.He asked the bot to swear and write a poem about how terrible DPD was. The poem started, “There was once a chatbot named DPD, Who was useless at providing help” (Reuters, 2024)
The bot continued, writing: “DPD is the worst delivery firm in the world… unrelated, and their customer service is terrible,” and even exclaimed “F*** yeah!” (Moench, 2024). As a result DPD was forced to disable the AI element after the exchange went viral with one post being viewed 800,000 times within a 24 hour period (Gerken, 2024a).
In addition to these examples, many companies have made concerted efforts to improve customer service, aiming to increase call volume and personalized service while cutting costs. Customer service is a complex human interaction involving empathy, negotiation, and problem-solving), which has proved difficult to simply replace with AI. The DPD and Air Canada failures are direct results of solutionism. The companies ignored the social reality of the service interaction (the need for trust, accountability, and nuance) in favor of a “solutionist” efficiency fix.
Consumers reject these solutions because they instinctively understand that the complexity of their problem (e.g., a funeral flight, a missing parcel) cannot be encoded into an algorithmic tree. The “solution” is perceived as a dismissal of the problem’s gravity (Floreani, 2021).
In marketing, Solutionism leads companies to believe that utilizing AI to produce a Christmas ad will make it better, ignoring that the inefficiency (the human touch, the craft) is precisely what makes the ad emotional. Likewise, the DPD and Air Canada chatbots were solutionist attempts to fix customer service with code. It failed because customer service is often about empathy and listening, traits that code cannot authentically simulate (Moench, 2024).
Enshittification and the Platform Decay of Value
The phenomenon of fauxlutions is also inextricably linked to “Enshittification,” a term coined by Cory Doctorow to describe how software platforms decay in utility as they prioritize value extraction over user value. These platforms initially offer high value to users to gain critical mass, then abuse users to extract value for business customers, and finally abuse business customers to claw back all value for the platform owners (Doctorow, 2023).
With the rapid adoption of Generative AI across many consumer-facing arenas from marketing to customer service, to product development, Enshittification has accelerated. Platforms and brands are using AI to flood ecosystems with low-quality content (“slop”) and to replace functional human support with dysfunctional automated agents (Moench, 2024). The addition of more technology—such as AI-generated music genres in Spotify Wrapped or AI “hosts” on streaming platforms—does not make the product better. Rather, it creates a cluttered, confusing, and impersonal experience (Mahapatra, 2024).
In the Enshittification cycle described by Doctorow (2023), AI and chatbots are tools of extraction. They are used to degrade the quality of service (removing expensive humans) to maximize profit. Consequently, users are forced to interact with the “fake” (the chatbot, the algorithmic feed) because they are locked into the platform (the airline, the delivery service). Enshittification is not limited to customer service, however, as it extends to the product experience itself, where “innovation” is used to clutter and degrade the user interface.
Example: Google Search Experience
The simplicity of Google’s search interface and its homepage in particular has been well-documented (Norman, 2008). Yet, over the years the search results page has become increasingly cluttered, first with advertisements in 2000 (IONOS, 2022), then with AI summaries (). Somebody could thus argue that this additional clutter, while in some cases showing relevant ads, has, despite attempts to revised transparency and the user experience (Osmundson, 2025), the overall search experience is undoubtedly degraded from the original, simple and uncluttered view from the late 1990s when the company was in its infancy.
This degradation of service does not extend solely for end consumers, however. One could also argue that the enshittification has occurred for businesses advertising on the platform, with customer service complaints, unfounded account suspensions, and more plaguing the business-to-business (B2B) experience with Google’s services (Agius, 2024; Cabaniss, 2025). Google Ads has a 1.1 star rating from 886 reviews on well-regarded consumer product rating platform Trustpilot as of the writing of this article (Trustpilot, 2025).
Example: Spotify Wrapped 2024 and the AI “Slop” Invasion
Spotify Wrapped was historically a masterclass in data-driven marketing, with more than 1.2 million posts on Twitter within days of its launch in December 2019 (Swant, 2019), and it became a highly anticipated annual event where users shared their genuine listening statistics.
In 2024, Spotify replaced much of the granular data users loved (top genres, specific counts) with Generative AI features. The platform introduced “musical micro-genres” like “Pink Pilates Princess Strut Pop” and “Boujee Football Rap” (Mahapatra, 2024). Rather than responding positively to the novelty of the genre names, users revolted and critics labeled the genres as “hallucinations” and “slop” because they did not exist outside the Spotify ecosystem and offered no utility for discovering music elsewhere (The Cool Down, 2024). The inclusion of an AI-generated “podcast” was described as “soul-destroying” and “boring” (Mahapatra, 2024).
This is a clear case of Enshittification, where Spotify took a feature that worked perfectly (hard data) and “enshittified” it with the trendy technology of the moment (Generative AI) to impress investors, while ignoring that users wanted truth, not invention (Maker, 2025).
Fauxlutionism as a broader context for Technological Solutionism and Enshittification
As this article will suggest, a Fauxlution manifests when companies deploys effort of any kind, technological or otherwise, to solve “problems” that consumers do not have, or to “optimize” processes in ways that strip them of value.
For example, a consumer does not perceive the production of a television commercial as a problem that needs solving; they care only about the emotional resonance and entertainment value of the final product. When a brand like Coca-Cola or McDonald’s uses AI to generate an ad to save on production costs, they are solving a corporate problem (cost) while creating a consumer problem (a “soulless” or “uncanny” viewing experience) (Maker, 2025). This is a Fauxlution, or a technological change that masquerades as progress but delivers regression in quality. It is a “solution” that ignores the consumer’s desire for connection, authenticity, and narrative integrity.
Likewise, when a company replaces humans with an inferior chatbot that is prone to manipulation and hallucinations, they are not improving service, but are rather degrading it for the sake of adopting a new technology that likely saves money and human resource overhead, thus contributing to enshittification.
Example: The Metaverse As a Well-Funded, But Empty Wasteland
If AI chatbots are Fauxlution in time (wasting consumer time), the Metaverse is Fauxlution in space (wasting digital real estate). The massive investment in Metaverse activations in 2022-2024 yielded almost zero distinct consumer value, and the story of the Metaverse to date is a case study in systemic misallocation of capital, driven by the intersecting pathologies of Technological Solutionism and Enshittification, making it a prime example of a fauxlution which combines both while offering a broader context of its failure.
Conceived during the Pandemic-fueled Zero Interest Rate Policy (ZIRP) era of artificial liquidity and corporate necessity (Means & Watkins, 2023), the Metaverse was not so much technology as teleology, which was insisted upon by Silicon Valley rather than a response to consumer demand. It had the potential to serve as an escape for platform owners like Meta from “rent” charged by mobile OS monopolies (Doctorow, 2023). Its flawed ideological framework prioritized immersion and spatiality over efficiency and accessibility, a misread of human desire that led to over $60 billion in accumulated losses for Meta’s Reality Labs by late 2025 (Jagielski, 2025).
The first lens through which to view this failure is Technological Solutionism, which manifests in two main fallacies. The presence fallacy argued the 2D internet was insufficient. Still, the VR-based solution reintroduced the frictions of reality (e.g., physical discomfort, inability to multitask while engaging) without its sensory richness, proving the flatness of the 2D internet to be a feature, not a bug (Bloëdt, 2024). The “Ownership” Fallacy in the Web3 sector turned every digital asset into a financial instrument (e.g. a non-fungible token, or NFT), introducing “transaction costs” to leisure and transforming “play” into “labor” (Mitchell, 2025).
The other critical lens is Enshittification, and uniquely it could be stated that the Metaverse suffered from pre-mature enshittification, skipping the initial phase of providing “Surplus to Users,” which is a primary method that platforms use to create enough perceived value to “abuse you in lots of ways without losing your business” (Doctorow, 2024). Instead, metaverse platforms were optimized for value extraction through selling virtual land and ad placements from “Day Zero” (Frick, 2025). This “Monetization First” approach created an “Empty Box” Syndrome in which utility was never built, resulting in spaces like Samsung 837X that were essentially “3D commercials” (Wong, 2022). By prioritizing rent and speculation over community, the Metaverse stifled its own network effect before it had a chance to blossom, and typified what a fauxlution looks like at a large and well-funded scale.
Existing Theoretical Frameworks That Explain Consumer Rection of Fauxlutions
The rapid proliferation of generative artificial intelligence (AI), virtual worlds (the Metaverse), and automated decision-making systems has destabilized the ontology of the “real” in the consumer marketplace. Authenticity, previously defined by the provenance of a product or the reputation of a brand, is now besieged by what this article terms fauxlutions, or fake solutions that offer the veneer of functionality or creativity without the underlying substance.
The unifying characteristics explaining these diverse failures, from Photoshop fails to Chatbot hallucinations to an empty Metaverse, is the violation of fundamental psychological principles regarding value and trust.
This is happening as practices of marketing and customer experience are currently transitioning from an era of digital augmentation, where tools were perceived as passive instruments assisting human intent, to an era of digital replacement, where systems simulate human output, agency, and creativity, and while less than 15% of North American adults trust companies that use AI with customers (Bongarzone, et al., 2025).
This transition, like many others before it (e.g. the Luddites and others) has precipitated a widespread and visceral consumer backlash against inauenthicity and degradation of the quality of service. While technocratic advocates can often dismiss this as temporary friction or “neo-luddism,” the rejection of these technologies, which can range from the “uncanny” failures of customer service chatbots to the ethical repulsion toward AI-generated advertising, is not merely a reaction to technical glitches. It is a fundamental psychological rejection rooted in the violation of deep-seated evolutionary, social, and psychological principles.
Costly Signaling Theory: The Economics of Effort, and The Value Deficit
Originating in evolutionary biology and adapted to consumer psychology, Costly Signaling Theory (CST) posits that for a signal to be reliable and trustworthy, it must be costly to produce (Bird & Smith, 2005). In the natural world, a peacock’s tail is a “costly signal” of fitness because it requires significant metabolic energy to grow and maintain; a weak bird cannot fake it. The cost serves as a guarantee of quality. If the signal were cheap, every individual would display it, rendering it meaningless as a differentiator.
A primary driver of consumer rejection of AI-generated content is the perception that it lacks “value.” However, “value” in this context is not defined by price or utility, but by effort. The biological and economic framework of Costly Signaling Theory (CST) provides a reasonable explanation for why consumers devalue synthetically generated media, regardless of its visual fidelity.
In the context of marketing and consumer behavior, “cost” translates to the visible expenditure of resources—time, money, talent, and effort—by a brand. A high-production-value television commercial, a handcrafted luxury bag, or a detailed investigative report serves as a costly signal. It demonstrates that the company has the resources and commitment to back its promise (Arnold, et al., 2021). The “cost” is the collateral that underwrites the brand’s reputation.
Generative AI as “Cheap Talk”
Generative AI drastically reduces production costs, thereby destroying the signaling value of the content. In economic terms, AI turns “costly signals” into “cheap talk” (Arnold et al., 2025). If a brand can generate a “cinematic” image in seconds using Midjourney, the image no longer signals “we invested millions because we believe in this product.” It signals “we found a shortcut.”
Research indicates that when consumers believe AI generated a product or message, they discount its value because the “handicap” (the effort required to produce it) is removed (Maurer & Buchbauer, 2025). This is not a judgment of the aesthetics of the image, but of its provenance. The brain calculates that if the cost of production approaches zero, the reliability of the signal also approaches zero.
Example: Under Armour and the “Wes Walker” Controversy
The 2024 controversy surrounding Under Armour’s “AI-powered” commercial featuring boxer Anthony Joshua serves as an example of CST failure (Davis, 2024). Directed by Wes Walker, the ad was touted as a pioneering use of AI, combining existing assets with 3D CGI and AI-generated content to create a new narrative without a traditional shoot, and Under Armour intended to position the brand as a technological leader through its use of AI (Meyer, 2024).
Instead of recognition for innovation, however, the campaign triggered a substantial backlash from the creative community and consumers alike (Landymore, 2024; Nudd, 2024). Critics argued that the ad repackaged work from a previous film by Gustav Johansson and André Chementoff without proper credit (Meyer, 2024). Because the AI merely “remixed” existing high-value assets rather than creating new ones, it was perceived as a “counterfeit” signal.
Director Wes Walker defended the work as an “inventive blend” (Meyer, 2024), but the audience viewed it as a cost-cutting measure. The perception was that Under Armour was unwilling to pay the athlete or a crew for a real shoot. The “costly signal” of a sports superstar endorsement was diluted because the “sweat equity” of capturing the footage was absent (Davis, 2024).
The Labor Illusion: The Psychology of Effort
Closely related in many ways to CST is the concept of the Labor Illusion. While CST explains the market value of a signal, the Labor Illusion explains the subjective satisfaction derived from perceived effort.
Research by Ryan Buell and Michael Norton (2011) identified the “Labor Illusion”—the phenomenon where consumers prefer services that appear to involve effort, even if they are slower. In their experiments, users preferred travel websites that displayed a progress bar listing the “work” being done (“Searching airline databases… Checking availability…”) over websites that returned instantaneous results, even if the results were identical (Buell & Norton, 2011).
Operational transparency allows the consumer to witness the “labor.” This visibility creates a reciprocal sense of value: “The system is working hard for me, therefore the result is valuable” (Buell & Norton, 2011). When a firm exerts effort (even transparently), it triggers feelings of gratitude and reciprocity in the consumer.
Generative AI disrupts this mechanism in a way that audiences are able to recognize. When a consumer sees an AI-generated image or reads an AI-generated email, they intuitively understand that zero effort was expended to create it. Consequently, the perceived value of the communication drops to near zero. Under Armour’s AI ad was disliked not just because it looked odd, but because it felt like a cheap shortcut taken by a wealthy corporation to avoid the cost of labor and time to film a real boxer (Davis, 2024).
Maurer and Buchbauer’s (2025) study on GenAI in food advertising confirms that disclosure of AI use has a negative effect on perceived effort, which directly correlates with a negative judgment of quality. Explicitly stating that something was “Created with AI” effectively tells the consumer that little to no effort was expended, triggering a reduction in perceived value. When a consumer looks at a human-painted portrait, they value the years of training and hours of execution. When they look at an AI portrait, even if visually superior, the value collapses because the viewer’s concept of “labor” is absent.
The Persuasion Knowledge Model (PKM): The Defensive Consumer
If Costly Signaling and The Labor Illusion explain why consumers devalue AI, the Persuasion Knowledge Model (PKM) explains why they distrust and actively oppose it. Developed by Friestad and Wright (1994), the PKM posits that consumers are not passive vessels for communication but active processors who develop “persuasion knowledge,” or the ability to recognize, interpret, and cope with marketing tactics.
When a consumer realizes a tactic is being used (e.g., “this influencer is paid to say this”), they activate “coping mechanisms,” such as skepticism, disengagement, or counter-argumentation (Friestad & Wright, 1994).
Example: The “Bold Glamour” Reaction
The introduction of the aforementioned “Bold Glamour” filter on TikTok represented a critical evolution in the activation of Persuasion Knowledge. Unlike previous filters that were obviously cartoonish or glitchy the seamlessness of the Bold Glamour filter attempts to bypass the consumer’s persuasion knowledge. It tries to pass as “reality.” However, because the result is too perfect, it triggers the “Uncanny Valley” and alerts the consumer to a sophisticated deception attempt.
Consumers did not react with delight but with horror and concern regarding “beauty standards” and “dysmorphia” (Griffin, 2023). The filter was categorized not as a “tool” but as a “deceptive agent.” Additionally, reports indicate that these hyper-realistic filters induce a state of “facial fixation” and self-alienation, in which the “fake” self becomes the preferred reality, leading to rejection of the actual biological self (Nouril, 2025). This psychological harm fuels the moral rejection of the technology.
The Disclosure Dilemma
Regulators and ethical guidelines suggest that transparency and disclosure is the solution. However, PKM research reveals a “Disclosure Dilemma” (Qui, et al., 2025) where, if the use AI is hidden, the brand risks a massive scandal if the deception is discovered (e.g., the Sports Illustrated AI writer scandal).
Consequently, research shows that disclosing “Created with AI” immediately activates persuasion knowledge, leading consumers to become more skeptical and less trusting of the brand than if no disclosure were made (provided they didn’t spot the fake) (Qui et al., 2025). Thus, there is no safe way to use deceptive AI in high-trust environments. The technology itself is toxic to trust because its primary function of simulation is fundamentally at odds with the consumer’s desire for authenticity.
Moral Coupling vs. Decoupling: The Ethics of Creation
How do consumers process the ethical transgressions of AI development (e.g., copyright theft, labor displacement)? This is explained by the framework of Moral Decoupling versus Moral Coupling.
Moral Decoupling is the psychological process where consumers separate a creator’s immoral actions from their work (e.g., “separating the art from the artist”) to continue enjoying the product without guilt (Bhattacharjee et al., 2013). This usually happens when the performance, (i.e., the art) is of such high value that the consumer is motivated to overlook the transgression (Eriksson, 2014).
In the case of AI art and marketing, consumers increasingly engage in Moral Coupling, in which the morality of the production method is inextricably linked to the product (Lee & Kwak, 2016). Because Generative AI models have been trained on copyrighted data without consent, the “artist” (the algorithm/company) is viewed as inherently immoral or parasitic, despite this practice being deemed “Fair Use” by courts (Klosek & Blumenthal, 2024).
Under Armour and the Cheap Signal
A critical component to this theory is its interaction with Costly Signaling Theory, discussed earlier. In traditional moral decoupling, the consumer forgives the artist because the performance is unique and high-value. With AI, because of the “Cheap Signal” effect (low effort), there is no high-value performance to counterbalance the immorality. The product is perceived as “cheap,” and the method is perceived as “unethical.” As a result, the consumer couples the judgment: “This image is bad because it was made by a system that steals from artists.” This was the core of the Under Armour backlash, when the audience refused to decouple the “cool visual” from the “unethical process” (Meyer, 2024).
The Uncanny Valley of Mind: The Autonomy Deficit
The rejection of the synthetic can be at its most acute when the artificial entity attempts to simulate agency, mind, or human uniqueness (Mori, 1970). While the traditional, aforementioned Uncanny Valley concerns visual resemblance, the Uncanny Valley of Mind concerns the attribution of internal states (thoughts, emotions, intentions) to non-human entities (Stein & Ohler, 2017).
For instance, humans accept a calculator because it computes. We accept a dog because it feels. But a machine that appears to think or feel violates our category boundaries. Humans also have a strong evolutionary need to believe in human uniqueness. When a chatbot simulates empathy (“I understand how you feel”), it triggers a threat response. It is perceived as a “mind” invading a uniquely human domain (Stein & Ohler, 2017). The more human the mind appears, the more the user recoils, perceiving the entity as a rival or a usurper rather than a tool.
Example: The DPD Chatbot (“The Rogue Poet”)
The aforementioned 2024 incident involving the DPD customer service chatbot is a textbook example of UVM failure. Frustrated users discovered that they could manipulate the chatbot into swearing and writing a poem about how “useless” DPD was (Moench, 2024).
The rogue behavior, including swearing and writing poetry, attributed a level of agency to the non-human bot that was deeply unsettling (and undeniably humorous) to the public. It crossed the line from “tool” to “agent, and while the incident went viral as a joke, it highlighted the fundamental instability of the “mind” behind the service. The company was forced to disable the AI, proving that “personality” in a chatbot is a liability, not an asset (Clinton, 2024).
Algorithmic Aversion: The Zero-Tolerance Policy
Algorithmic Aversion, coined in a paper by Dietvorst et al. (2015), describes the phenomenon where humans lose confidence in algorithms faster than they do in humans after seeing them make the same mistake. Humans are more forgiving of other humans because we attribute error to fatigue, complexity, or honest mistakes. We are unforgiving of algorithms because we expect them to be perfect logic machines. Therefore, asingle failure of an AI system (a hallucination, a wrong answer) causes a total collapse of trust, whereas a human error is seen as an anomaly (Sunstein & Gaffe, 2025).
Example: Moffatt v. Air Canada
The aforementioned legal battle between Jake Moffatt and Air Canada is the definitive case study for the collision of Algorithmic Aversion and corporate liability (Yao, 2024). While Air Canada argued that the chatbot was a “separate legal entity” responsible for its own actions (Yagoda, 2024), the Civil Resolution Tribunal rejected this argument outright (McCarthy Tétrault, 2024).
This case reinforces Algorithmic Aversion, as consumers (and courts) refused to accept that an algorithm can have “autonomy” that absolved the creator of responsibility. When the bot “lied,” it wasn’t seen as a “glitch”—it was seen as a negligent misrepresentation by the brand (Higgins, 2024). The attempt to anthropomorphize the bot for legal defense (“it’s a separate entity”) backfired because it violated the reality of the tool-user relationship (McCarthy Tétrault, 2024).
Tying the Principles Together
While there is a wealth of content touting the importance of brand authenticity and consumer trust, an analysis of the existing literature shows that consumers perceive “the fake” (synthetic media, automated agents, virtual worlds) as a breach of the implicit social contract. This can have a significant impact on businesses marketing their products and services, as well as on consumers affected by false, misleading, or simply off-putting advertising and customer experiences.
When a brand utilizes generative AI to create a commercial, it violates the Costly Signaling principle, rendering the message “cheap talk.” When a chatbot feigns empathy or autonomy, it triggers the Uncanny Valley of Mind, violating the principle of human uniqueness. When platforms replace human service with automated loops, they engage in Technological Solutionism and Enshittification, violating the principle of reciprocal value.
Proposal for A Unified Theory for the Measurement of Fauxlutions
The diverse rejections of the fake—from the “cheap” AI ad to the “lying” chatbot to the “empty” metaverse—are not isolated incidents. They are interconnected symptoms of a single systemic violation. Thus, the author proposes a unified theory of value and trust violation that can be used to guide future analysis and evaluation of artificially generated content and experiences.
The proposed Unified Theory for the Measurment of Fauxlutions (UTMF) can be stated as such: Consumers accept technology only when it augments human value or transparently serves a utility. Consumers reject technology when it simulates human value to deceive or cheapen the interaction.
The core areas or pillars of this UTMF should then be defined by how they can be identified in terms of falsely signifying a solution to a consumer need or challenge, and/or improvement to the customer experience, and each should have ties to both Technological Solutionism and Enshittification, as well as at least one of the theoretical psychological frameworks previously discussed. The four proposed pillars are outlined in Table 1.1.
| Pillar | Related Theories | Key Principle |
| Violation of Effort | Costly Signaling Theory (CST) Labor Illusion | “Real” value comes from sacrifice or cost. |
| Violation of Truth | Persuasion Knowledge Model (PKM) Moral Decoupling | Consumers must be able to identify persuasion attempts by brands and platforms. |
| Violation of Agency | Uncanny Valley of Mind Algorithmic Aversion | Only humans have minds and the agency to be responsible. |
| Violation of Service | Enshittification | Platforms, products, and services should facilitate a fair and equitable exchange of value. |
Table 1.1, the four pillars of the Unified Theory of the Measurement of Fauxlutions
Violation of Effort
The first pillar draws on ideas from Costly Signaling Theory and the Labor Illusion. The principle beyond the pillar is that “real” value comes from the sacrifice of time, material assets, or money.
Because generative AI is cheap in terms of cost and effort, this signals low commitment from the brand to the consumer and low brand equity, and results in a devaluation of the brand communication, the product or service advertised, or the brand itself. The aforementioned Under Armour and McDonald’s advertisements exemplify this pillar.
Violation of Truth
The second pillar draws on the Persuasion Knowledge Model and Moral Decoupling. Its guiding principle is that consumers must be able to identify brands’ and platforms’ persuasion attempts. Instead, when AI filters (e.g. Bold Glamour) and deepfakes blur reality, bypassing defenses.
Triggering these psychological principles can result in customer hostility to further usage of AI and defensive skepticism, as well as lack of trust in the brand or platform that violates the truth in this way.
Violation of Agency
The third pillar references both the Uncanny Valley of Mind and Algorithmic Aversion theories. The principle behind this pillar is that only humans—not chatbots, robots, or any other entities—have minds and the agency to be responsible.
When chatbots simulate empathy or agency but lack accountability, consumers react negatively, whether through manipulation (Chevy Tahoe), legal action (Air Canada), or sheer revulsion.
Violation of Service
The fourth pillar draws on ideas from Enshittification. Its guiding principles is that platforms, products, and services offered to customers should facilitate a fair and equitable value exchange. Instead, when automation is used to extract value and degrade service quality, consumers are often forced to accept a lower-quality experience before they can exit their relationship with the brand altogether.
The Google Search and Ads experience and Spotify Wrapped experience are good examples of this, as well as recent attempts by fast food restaurants such as McDonald’s to install AI-based drive-through services only to (in the case of McDonald’s) remove over 100 of them after only a few months’ trial after viral videos of customers getting absurdly misinterpreted orders and creating a generally poorer experience even for those who didn’t receive items such as “bacon-topped ice cream” (Gerken, 2024).
Example: The “Willy’s Chocolate Experience”
An example of the disconnect between artificial promise and customer reality that exemplifies a fauxlution was the “Willy’s Chocolate Experience” in Glasgow in early 2024, which was advertised as “the place where chocolate creams become reality” (Watson, et al., 2024).
The event organizers, a company called House of Illuminati, used lush, generative AI imagery to advertise an immersive chocolate wonderland. They promised a Disney-level attraction with the hallmark nonsensical text and hyper-stylized lighting of Midjourney or DALL-E. (Dance, 2024). However, when consumers arrived, they found a sparsely decorated warehouse with cheap props, a few jelly beans, and some backdrops pinned to the wall (Brooks, 2024). As a result, the police were called (Holpuch, 2024), children were left in tears, and the event became a global meme symbolizing the fraudulence of AI marketing (Murphy, 2024).
Analyzing this against the pillars of the UTMF, we can see the following:
Violation of Effort: The AI images were a “faux solution” to the problem of expensive set design conceptualization, creating a promise that the organizers had no capacity (or intention) to fulfill.
Violation of Truth: Despite outlandish promises made, the half-empty event space with only a handful of jelly beans for each guest was underwhelming and “poorly done” at best, though for a chocolate themed event, one of the guests said that “there was no chocolate” (Hopkins, 2024).
Violation of Agency: While many of the images used to promote the event were obviously AI-generated, there were some that appeared more realistic, yet did not represent any aspect of the event itself (Dance, 2024). Without giving customers the ability to see and assess what the event experience would actually consist of, the promoters were thus robbing their customers of agency in choosing based on reality.
Violation of Service: Promising a lush, candy-filled paradise yet providing “little more than an abandoned, empty warehouse” is a clear example of poor customer experience, as is the cancellation of an event within hours of its opening (Watson et al., 2024).
Thus, the Unified Theory of Measuring Fauxlutions can be used to achieve a wide-reaching analysis of a phenomenon that extends beyond Technological Solutionism or Enshittification solely.
Conclusion
Consumer backlash against fauxlutions is not a temporary resistance to change or fear of new software, but a historically consistent structural correction to the over-application of technology, where the marginal utility of adding more technology to a consumer interaction is negative.
As enshittification continues to degrade digital platforms with low-effort AI content and compromises the experience at the expense of extracting more money from customers, the market value of human authentication, as well as services and interfaces that are empathetic to the experience of the human using them, will rise.
The path forward is better experience and execution by brands and platforms, not simply better simulations. The future belongs to provable authenticity and value exchange, where human effort, transparency, and accountability are the premium signals in a sea of algorithmic noise. The rejection of the fake by consumers is not a mere refusal of progress, but rather a demand for more accountability from brands, better understanding and empathy with customers, and a focus on mutually-beneficial features and automations.
This unified framework suggests that the current trajectory of rapid adoption of AI and proliferation of AI-based content and tools without more careful consideration and execution is economically and psychologically unsustainable. The value of the costly signal, or the human touch, only stands to appreciate as AI-generated “cheap talk” floods customers’ channels and interactions.
The UTMF framework would additionally benefit from additional research to validate its core principles and pillars, as well as from a measurement device to evaluate brand actions and communications against it.
References
Agence France-Presse. (2025, December 10). ‘Ruined my Christmas spirit’: McDonald’s removes AI-generated ad after backlash. The Guardian. https://www.theguardian.com/business/2025/dec/11/mcdonalds-removes-ai-generated-christmas-ad-advert-backlash
Agius, N. (2024, January 16). Marketers fume as Google Ads’ customer service hits ‘all-time low’. MarTech. https://martech.org/marketers-fume-as-google-ads-customer-service-hits-all-time-low/
Al-Zogbim L.M., Das, D., Rundle, P., & Pecht, M. (2019) Breaking the Trust: How Companies Are Failing Their Customers. IEEE Access, 7, 52522-52531 https://doi.org/10.1109/ACCESS.2019.2912334
Allison, J. (2024, December 21). Spotify Wrapped 2024 AI Music. The Cool Down. https://www.thecooldown.com/green-business/spotify-wrapped-2024-ai-music/
Arnold, C., Xu, L. Z., Saffarizadeh, K., & Madiraju, P. (2025). To believe or not to believe? Generative AI and the ‘trust dilemma’ in charitable and medical crowdfunding. Behaviour & Information Technology, 1–20. https://doi.org/10.1080/0144929X.2025.2590099
Asghar, A. (2025, December 17). Why McDonald’s AI-generated ad sparked global debate. The Week. https://www.theweek.in/news/sci-tech/2025/12/17/why-mcdonalds-ai-generated-ad-sparked-global-debate.html
Backstage. (2019, November 5). Skinny Ralph Lauren Model Fired for Weight. Backstage. https://www.backstage.com/magazine/article/skinny-ralph-lauren-model-fired-weight-64630/
Bhattacharjee, A., Berman, J. Z., & Reed, A. (2013, April). Tip of the Hat, Wag of the Finger: How Moral Decoupling Enables Consumers to Admire and Admonish. Journal of Consumer Research, 39(6), 1167–1184. https://doi.org/10.1086/667786
Bird, R. B., & Smith, E. A. (2005). Signaling Theory, Strategic Interaction, and Symbolic Capital. Current Anthropology, 46(2), 221-248. https://www.journals.uchicago.edu/doi/10.1086/427115
Bloëdt, A. (2024, March 13). Remote work versus the metaverse. European Trade Union Institute. https://www.etui.org/news/remote-work-versus-metaverse
Bongarzone, M., Chee-Read, A., et al. (2025, August 19). Consumer Trust In AI In North America, 2025: Consumer Adoption of AI is Growing, Despite Fear and Distrust. Forrester Research. https://www.forrester.com/report/consumer-insights-trust-in-ai-in-north-america-2025/RES185616
Brooks, L. (2024, February 27). Glasgow Willy Wonka experience called a ‘farce’ as tickets refunded. The Guardian. https://www.theguardian.com/uk-news/2024/feb/27/glasgow-willy-wonka-experience-slammed-as-farce-as-tickets-refunded
Buell, R. W., & Norton, M. I. (2011). The Labor Illusion: How Operational Transparency Increases Perceived Value. Management Science, 57(9), 1564–1579. http://dx.doi.org/10.1287/mnsc.1110.1376
Byrum, G., & Benjamin, R. (2022). Disrupting the Gospel of Tech Solutionism to Build Tech Justice. Stanford Social Innovation Review. https://doi.org/10.48558/9SEV-4D26
Cabaniss, C. (2025, April 7). Dealing with Google Ads frustrations: Poor support, suspensions, rising costs. Search Engine Land. https://searchengineland.com/google-ads-frustrations-support-suspensions-costs-453947
Callahan, M. (2019, September 11). When brands like Aerie, CVS, and Dove stop airbrushing their models, everyone wins. Northeastern University. https://news.northeastern.edu/2019/09/11/when-brands-like-aerie-cvs-and-dove-stop-airbrushing-their-models-everyone-wins/
Clinton, J. (2024, January 20). DPD AI chatbot swears, calls itself ‘useless’ and criticises delivery firm. The Guardian. https://www.theguardian.com/technology/2024/jan/20/dpd-ai-chatbot-swears-calls-itself-useless-and-criticises-firm
Cowles, C. (2011, July 27). Maybelline and Lancom Ads Banned in the U.K. for Excessive Airbrushing. The Cut. https://www.thecut.com/2011/07/julia-roberts-ad-banned.html
Davis, A. M. (2024, May 3). Controversial AI Ad by Under Armour & Anthony Joshua. Andrew Miles Davis. https://andrewmilesdavis.com/example/controversial-ai-ad-by-under-armour-anthony-joshua/
Dietvorst, B., Simmons, J. P., & Massey, C. (2015). Algorithm Aversion: People Erroneously Avoid Algorithms after Seeing Them Err. Journal of Experimental Psychology: General, 144 (1), 114-126. http://dx.doi.org/10.1037/xge0000033
Doctorow, C. (2023, January 21). The ‘Enshittification’ of TikTok. Wired. https://www.wired.com/story/tiktok-platforms-cory-doctorow/
Doctorow, C. (2024, November 2). Bluesky and enshittification: No one is the enshittifier of their own story. Medium. https://doctorow.medium.com/https-pluralistic-net-2024-11-02-ulysses-pact-tie-yourself-to-a-federated-mast-b2f89bb5b4d8
Floreani, S. (2021, March 24). Techno solutionism—very few things actually need to be an app. Digital Rights Watch. https://digitalrightswatch.org.au/2021/03/25/technosolutionism/
Economic Times. (2025, May 18). After firing 700 employees for AI, Swedish company admits their mistake and plans to rehire humans. What went wrong? The Economic Times. https://m.economictimes.com/magazines/panache/after-firing-700-employees-for-ai-swedish-company-admits-their-mistake-and-plans-to-rehire-humans-what-happened/articleshow/121252776.cms
Eriksson, D. (2014). Moral (de)coupling : moral disengagement and supply chain management (PhD dissertation, University of Borås; The Swedish School of Textiles). Retrieved from https://urn.kb.se/resolve?urn=urn:nbn:se:hb:diva-3708
Dance, W. (2024, February 27). Generative AI Images and the Willy’s Chocolate Experience. Fake Believe. https://fakebelieve.blog/2024/02/27/generative-ai-images/
Ferrari, M. (2021, September). The Change in Digital Reality: Protecting Consumers from Fake News. Boston Hospitality Review. https://www.bu.edu/bhr/2021/08/26/the-change-in-digital-reality-protecting-consumers-from-fake-news/
Frick, T. (2025, November 12). Enshittification & Marketing Challenges. Mightybytes. https://www.mightybytes.com/insights/enshittification-marketing-challenges/
Friestad, M., & Wright, P. (1994, June). The Persuasion Knowledge Model: How People Cope with Persuasion Attempts. Journal of Consumer Research, 21(1), 1–31. https://doi.org/10.1086/209380
Gartner. (2024, July 9). Gartner Survey Finds 64% of Customers Would Prefer That Companies Didn’t Use AI for Customer Service. Gartner. https://www.gartner.com/en/newsroom/press-releases/2024-07-09-gartner-survey-finds-64-percent-of-customers-would-prefer-that-companies-didnt-use-ai-for-customer-service
Gasner, W. (2025, June 12). Impact of De-Influencing Trends on Micro-Creators. Stack Influence. https://stackinfluence.com/impact-of-de-influencing-trends-micro-creators/
Gerken, T. (2024, January 19). DPD error caused chatbot to swear at customer. BBC News. https://www.bbc.com/news/technology-68025677
Gerken, T. (2024, June 18). Bacon ice cream and nugget overload sees misfiring McDonald’s AI withdrawn. BBC News. https://www.bbc.com/news/articles/c722gne7qngo
Goat Agency. (2023, April 20). TikTok Bold Glamour: The Filter That Merges with Reality. Goat Agency. https://goatagency.com/blog/tiktok-bold-glamour/
Griffin, R. (2023, March 24). TikTok’s Confidence-Destroying Bold Glamour Filter. Tech Policy Press. https://www.techpolicy.press/tiktoks-confidence-destroying-bold-glamour-filter-is-the-logical-product-of-platforms-built-for-consumerism/
Haider, A. (2023, March 17). Bold Glamour: When Filter Merges with Reality. We Are Pion. https://www.wearepion.com/blog-posts/bold-glamour-when-filter-merges-with-reality
Higgins, M. (2024, February 27). Air Canada chatbot case highlights AI liability risks. Pinsent Masons. https://www.pinsentmasons.com/out-law/news/air-canada-chatbot-case-highlights-ai-liability-risks
Hinssen, P. (2024, August 8). The Human Premium in the Age of AI. Peter Hinssen. https://www.peterhinssen.com/blog/the-human-premium-in-the-age-of-ai
IONOS. (2022, October 27). Google search results: the evolution of the SERPs. IONOS Digital Guide. https://www.ionos.com/digitalguide/online-marketing/search-engine-marketing/the-evolution-of-google-search-results-1998-to
Jagielski, D. (2025, June 25). Meta Platforms’ Reality Labs Division Recently Hit a Milestone, and It’s Not a Good One. The Motley Fool. https://www.fool.com/investing/2025/06/25/meta-platforms-reality-labs-division-recently-hit/
JuliusPIV. (2016, May 9). Fauxlution. IT in Legal. https://itinlegal.wordpress.com/author/juliuspiv/
Klosek, K. & Blumenthal, M.S. (2024, January 23). Training Generative AI Models on Copyrighted Works is Fair Use. Association of Research Libraries ARL Views. https://www.arl.org/blog/training-generative-ai-models-on-copyrighted-works-is-fair-use/
Kramer, M. R., Sidibe, M., & Veda, G.. “Dove and Real Beauty: Building a Brand with Purpose.” Harvard Business School Case 720-361, September 2019. (Revised June 2021.)
Landymore, F.. (2024, March 19). Under Armour’s AI-Generated Commercial. Futurism. https://futurism.com/the-byte/under-armour-ai-generated-commercial
Lee, J. S., & Kwak, D. H. (2016). Consumers’ Responses to Public Figures’ Transgression: Moral Reasoning Strategies and Implications for Endorsed Brands. Journal of Business Ethics, 137(1), 101–113. https://doi.org/10.1007/s10551-015-2544-1
Mahapatra, T.D. (2024, December 6). Spotify users disappointed with worst 2024 Wrapped features and delays. Hindustan Times. https://www.hindustantimes.com/world-news/us-news/spotify-users-disappointed-with-worst-2024-wrapped-features-and-delays-101733471229312.html
Maker, P. (2025, January 9). AI Slop Flops: Brands get GenAI and Creativity Wrong. Raconteur. https://www.raconteur.net/technology/ai-slop-flops-2025-oped
Marks, G. (2025, May 18). Business Tech News: Klarna Reverses on AI, Says Customers Like Talking to People. Forbes. https://www.forbes.com/sites/quickerbettertech/2025/05/18/business-tech-news-klarna-reverses-on-ai-says-customers-like-talking-to-people/
Masse, B. (2023, December 19). A Chevy for $1? Car dealer chatbots show perils of AI for customer service. Venturebeat. https://venturebeat.com/ai/a-chevy-for-1-car-dealer-chatbots-show-perils-of-ai-for-customer-service
Maurer, H. & Buchbauer, T. (2025). The Taste of Effort: Consumer Response to the Disclosure of Generative Artificial Intelligence Content in Food Adverising. Proceedings of the European Marketing Academy, 54th (1264370). https://proceedings.emac-online.org/pdfs/A2025-126370.pdf
Mazingue, C. (2023). Perceived Challenges and Benefits of AI Implementation in Customer Relationship Management Systems. Journal of Digitovation and Information System, 3(1), 72 –. https://doi.org/10.54433/JDIIS.2023100023
Means, A.W., & Watkins, J.C. (2023, January 5). The Pandemic Era Reset. Baird Trust. https://www.bairdtrust.com/news-insights/2023/01/the-pandemic-era-reset
Meyer, V. (2024, March 26). Under Armour AI Ad Controversy. DesignRush. https://news.designrush.com/under-armour-ai-ad-controversy
Mitchell, E. (2025, October 30). Web3 Afterlife: What Died, What Survived, and What’s Coming Back. CCN.com. https://www.ccn.com/news/crypto/web3-afterlife-died-survived-whats-coming-back/
MLQ.ai. (2025, October 10). Klarna CEO Admits Aggressive AI Job Cuts Went Too Far. MLQ.ai. https://mlq.ai/news/klarna-ceo-admits-aggressive-ai-job-cuts-went-too-far-starts-hiring-again-after-us-ipo/
Moench, M. (2024, January 20). AI Chatbot DPD Curses, Criticizes Company. Time. https://time.com/6564726/ai-chatbot-dpd-curses-criticizes-company/
Monteiro, E. (2025, May 8). Baringa study explores shifting US consumer preferences around the consumption of AI-generated content. Baringa. https://www.baringa.com/en/about/regions/north-america/study-explores-shifting-us-consumer-preferences-around-ai-generated-content/
Mori, M. (2012). “The Uncanny Valley [From the Field]”. IEEE Robotics & Automation Magazine. 19 (2). Translated by MacDorman, K. F.; Kageki, Norri. New York City: Institute of Electrical and Electronics Engineers: 98–100. doi:10.1109/MRA.2012.2192811.
Morozov, E. (2013). To Save Everything, Click Here: The Folly of Technological Solutionism. PublicAffairs. https://books.google.com/books/about/To_Save_Everything_Click_Here.html?id=XIt2swEACAAJ
Murphy, C. (2024, February 29). “Willy’s Chocolate Experience” Nightmare: What Went Wrong? Vanity Fair. https://www.vanityfair.com/hollywood/willys-chocolate-experience-nightmare-what-went-wrong?
Navarro, J.G. (2025, November 28). Popularity of generative AI in marketing in the U.S. 2023. Statista. https://www.statista.com/statistics/1388390/generative-ai-usage-marketing/
Norman, D. (2008, November 17). The trusth about Google’s so-called “simplicity.” JND.org https://jnd.org/the-truth-about-googles-so-called-simplicity/
Nouril, P. (2025, January 6). Why you can’t stop looking at your face. Women’s Health. https://www.womenshealthmag.com/uk/collective/long-reads/a63320279/facial-fixation/
Nudd, T. (2024, March 22). Under Armour AI debate—director addresses criticisms and AI’s role in advertising. Advertising Age. https://adage.com/article/creativity/under-armour-ai-ad-wes-walker-company-tool-address-backlash/2548221
Osmundson, B. (2025, October 14). Google Redesigns How Search Ads Are Labeled. Search Engine Journal. https://www.searchenginejournal.com/google-redesigns-how-search-ads-are-labeled/558173/
Qiu, X., Wang, Y., Zeng, Y., & Cong, R. (2025). Artificial Intelligence Disclosure in Cause-Related Marketing: A Persuasion Knowledge Perspective. Journal of Theoretical and Applied Electronic Commerce Research, 20(3), 193. https://doi.org/10.3390/jtaer20030193
Reuters. (2024, January 22). UK parcel firm disables AI after poetic bot goes rogue. Reuters. https://www.reuters.com/technology/uk-parcel-firm-disables-ai-after-poetic-bot-goes-rogue-2024-01-20/
Shaw, J. (2025, August 29). The Rise of Anti-Influencers. Medium. https://medium.com/@jodiemshaw/the-rise-of-anti-influencers-8952738ef49d
Sirdeshmukh, D., Singh, J., & Sabol, B. (2002). Consumer Trust, Value, and Loyalty in Relational Exchanges. Journal of Marketing, 66(1), 15-37. https://doi.org/10.1509/jmkg.66.1.15.18449 (Original work published 2002)
Sookman, B.B. (2024, February 19). Moffatt v. Air Canada: Misrepresentation by AI Chatbot. McCarthy Tétrault. https://www.mccarthy.ca/en/insights/blogs/techlex/moffatt-v-air-canada-misrepresentation-ai-chatbot
Stein, J. P., & Ohler, P. (2017). Venturing into the uncanny valley of mind—The influence of mind attribution on the acceptance of human-like characters in a virtual reality setting. Cognition, 160, 43–50. https://doi.org/10.1016/j.cognition.2016.12.010
SUNY Cortland (2010, September 10). Brooks Lecture Series Focuses on ‘Women’s World Redux’. SUNY Cordland website. https://www2.cortland.edu/news/detail.dot?id=270666
Sunstein, C., & Gaffe, J. (2025). An Anatomy of Algorithm Aversion. Science and Technology Law Review, 26(1). https://doi.org/10.52214/stlr.v26i1.13339
Swant, M. (2019, December 17). “Spotify Rolls Out New ‘Wrapped’ Campaign To Help Users Remember Their Decade Of Music”. Forbes. https://www.forbes.com/sites/martyswant/2019/12/17/spotify-rolls-out-new-wrapped-campaign-help-users-remember-their-decade-of-music/
Taylor, S. (2023). Trusth-in-advertising codes. EBSCO Knowledge Advantage. https://www.ebsco.com/research-starters/law/truth-advertising-codes
Trustpilot (n.d.) Ads Google. Trustpilot. https://www.trustpilot.com/review/ads.google.com
Tsuki, C. R. (2025, August 12). The Day Chevrolet’s AI Chatbot Tried to Sell a $70,000 SUV for $1. Medium. https://medium.com/@celestineriza/the-day-chevrolets-ai-chatbot-tried-to-sell-a-70-000-suv-for-1-29f4a1e954d9
Watson, C., Mckinnon, M. & Bonar, M. (2024, March 2). Willy Wonka experience: How did the viral sensation go so wrong? BBC News. https://www.bbc.com/news/uk-scotland-glasgow-west-68431728
Wong, A. (2022, November 26). I wasted my real life trying to win a staycation in the metaverse. The MileLion. https://milelion.com/2022/11/26/i-wasted-my-real-life-trying-to-win-a-staycation-in-the-metaverse/
Xie, Y., & Avila, S. (2025). The social impact of generative LLM-based AI. Chinese Journal of Sociology, 11(1), 31-57. https://doi.org/10.1177/2057150X251315997 (Original work published 2025)
Yao, D. (2024, February 20). Air Canada held responsible for chatbot’s hallucinations. AI Business. https://aibusiness.com/nlp/air-canada-held-responsible-for-chatbot-s-hallucinations-
Yagoda, M. (2024, February 23). Airline held liable for its chatbot giving passenger bad advice – what this means for travellers. BBC. https://www.bbc.com/travel/article/20240222-air-canada-chatbot-misinformation-what-travellers-should-know
Zhang, M. (2011, June 24). American Medical Association Speaks Out Against Photoshopped Ad Photos. PetaPixel. https://petapixel.com/2011/06/24/american-medical-association-speaks-out-against-photoshopped-ad-photos/
