# AI washing

AI washing is a deceptive marketing tactic in which companies overstate the role of artificial intelligence in their products or services to mislead consumers and investors. The term, first defined by the AI Now Institute in 2019, raises concerns about transparency, ethics, and legal compliance.

AI washing is a deceptive marketing practice in which companies promote a product or service by overstating the role of artificial intelligence (AI) and its integration into the design, development, or functionality of the offering. The tactic is used to mislead customers into perceiving enhanced value and to attract investment from venture capitalists and other financial backers. The practice has drawn scrutiny for its implications regarding transparency, adherence to security regulations, and the erosion of consumer trust in the broader AI industry. The term was first defined by the AI Now Institute, a research institute at New York University, in 2019, and it draws an intentional parallel to greenwashing, which similarly misrepresents a product's environmental impact.

The definition of AI washing has evolved beyond product marketing. In 2026, The Washington Post described AI washing as a trend in which corporate leaders attribute layoffs to AI-driven productivity gains, even when the workforce reductions may have been motivated primarily by cost-cutting or efforts to improve profitability. This expanded usage reflects a broader pattern of leveraging the AI label for reputational or financial advantage, regardless of the actual role of technology in corporate decisions.

## Origins and Etymology

AI washing was coined as a direct analog to greenwashing, a term that emerged in the 1980s to describe companies that publicly overstate their environmental friendliness while concealing harmful practices. The AI Now Institute introduced the term in 2019 in a report that analyzed the gap between corporate claims of AI adoption and the actual state of AI systems. The institute's research highlighted instances where companies labeled routine software features as 'AI-powered' to capitalize on the excitement surrounding [machine-learning](https://www.wikiprompt.org/wiki/machine-learning) and [deep-learning](https://www.wikiprompt.org/wiki/deep-learning) advances.

The linguistic construction - adding '-washing' to a technology or policy area - had already been used for other practices such as 'pinkwashing' and 'sports-washing,' where a superficial association with a cause or activity is used to obscure underlying issues. In the context of AI, the suffix signals that a company is engaging in a superficial or misleading association with the technology.

## Characteristics and Methods

AI washing appears in several common forms. A company may claim to have used [artificial-intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence) in the development of a product without any actual involvement of AI tools or techniques. Alternatively, a product may be marketed with buzzwords such as 'smart,' 'intelligent,' or 'AI-powered' even when it does not incorporate [neural-network](https://www.wikiprompt.org/wiki/neural-network) or [large-language-model](https://www.wikiprompt.org/wiki/large-language-model) technology in any meaningful way. Some companies overstate the sophistication of the AI they do use, for example, calling a simple rule-based system a [generative-ai](https://www.wikiprompt.org/wiki/generative-ai) solution.

The practice is frequently driven by competitive pressure - when rivals begin advertising AI features, other companies feel compelled to make similar claims to avoid appearing outdated. It can also be a deliberate strategy to secure funding from investors who are actively seeking exposure to the AI boom. In this sense, AI washing has been compared to the dot-com bubble of the late 1990s, when businesses appended 'dot-com' to their names to inflate their valuations without having viable internet business models.

## Regulatory and Legal Landscape

The United States Securities and Exchange Commission (SEC) has taken a direct interest in AI washing, viewing it as a form of securities fraud when it misleads investors. In March 2024, the SEC imposed its first civil penalties on two companies for making misleading statements about their use of AI. These penalties signaled that regulators would treat unsubstantiated AI claims as actionable. In July 2024, the SEC charged a corporate executive from an AI hiring startup with fraud related to the use of AI buzzwords in investor communications.

The regulatory approach is grounded in existing rules against deceptive marketing and false advertising. It also aligns with broader concerns about accountability in the AI industry, as discussed by researchers from institutions such as the [stanford-ai-lab](https://www.wikiprompt.org/wiki/stanford-ai-lab) and the [berkeley-ai-research](https://www.wikiprompt.org/wiki/berkeley-ai-research) group. The lack of standard definitions for what constitutes 'real' AI complicates enforcement, because companies can argue that their use of pattern recognition or automated decision-making qualifies as AI under a broad interpretation.

## Case Studies and Examples

In September 2023, Coca-Cola released a new product called Coca-Cola Y3000 and stated that the flavor had been 'co-created with human and artificial intelligence.' The company was widely accused of AI washing because it provided no concrete evidence of AI involvement in the flavor's development. Critics argued that AI was used primarily as a marketing tool to capture consumer attention, not as an integral part of product creation. The episode became a frequently cited instance of AI washing due to the brand's prominence and the vagueness of the claim.

A separate pattern emerged in 2026, when several technology companies faced accusations of AI washing for attributing layoffs to AI-driven productivity gains. In these cases, executives claimed that AI had made certain roles redundant and that workforce reductions were a natural consequence of automation. Investigative reporting suggested that the layoffs were more closely tied to cost-cutting measures and efforts to improve quarterly earnings. The gap between the public narrative and internal motivations became a new dimension of AI washing, moving the concept beyond product marketing into corporate communications.

## Impact on Consumers and Investors

AI washing deceives consumers by creating false expectations about product capabilities and by obscuring the actual mechanisms behind a service. Customers who purchase a product because it is labeled 'AI-powered' may discover that it functions no differently from a conventional alternative. This can lead to dissatisfaction, erosion of brand trust, and a general skepticism toward legitimate AI products.

Investors are similarly affected. Pitching a startup as an AI company when it lacks proprietary technology can distort capital allocation and inflate valuations. The AI ecosystem - including firms like [openai](https://www.wikiprompt.org/wiki/openai), [anthropic](https://www.wikiprompt.org/wiki/anthropic), and [google-deepmind](https://www.wikiprompt.org/wiki/google-deepmind) - depends on confidence in the applicability of AI to real-world problems. Overblown claims from less scrupulous companies can create an environment reminiscent of the dot-com crash, where public disappointment followed a period of irrational exuberance.

The phenomenon also hampers legitimate advancement in AI. When consumers and regulators become wary of AI claims, serious researchers and developers must invest additional effort in proving the efficacy of their systems. This dynamic can slow adoption of valuable tools such as [transformer](https://www.wikiprompt.org/wiki/transformer)-based language models or computer vision systems that have been validated through rigorous evaluation.

## Mitigation Strategies

Companies are expected to be transparent about the specific AI techniques they employ and to provide evidence of how those techniques contribute to the product's function. Publishing technical documentation, disclosing model architectures, and citing evaluation results can all reduce the risk of AI washing accusations. For example, a company that uses [fine-tuning](https://www.wikiprompt.org/wiki/fine-tuning) on a pretrained large language model should describe the training process and the tasks the model was optimized for, rather than simply slapping an 'AI' label on the product.

Consumers can mitigate the effects of AI washing by requesting hard evidence from companies regarding the usage of AI tools. Asking for case studies, performance benchmarks, or demonstrations of the AI system in action can help separate genuine integration from marketing gloss. Consumers should also evaluate the product or service as a whole rather than being swayed by the presence of AI terminology. Factors such as usability, reliability, and cost should remain central to the purchasing decision.

Regulators have an ongoing role in setting expectations. The SEC's enforcement actions in 2024 established a precedent, but the evolving definition of AI washing - including the 2026 interpretation related to layoffs - suggests that additional guidance may be needed. Industry groups and academic institutions, including the [ai-now-institute](https://www.wikiprompt.org/wiki/ai-now-institute), have called for clearer standards around AI disclosure.

## Relationship to Other Phenomena

AI washing is conceptually adjacent to the AI bubble, a term used to describe the potential overvaluation of AI companies based on unrealistic expectations for near-term returns. While the AI bubble refers to market dynamics at a macroeconomic scale, AI washing is a micro-level practice that can feed the bubble by encouraging speculative investment. The term vaporware is also related, as it refers to products that are announced or advertised long before they are actually delivered. AI washing can be seen as a variant of vaporware where the product exists, but the AI component that is advertised does not.

The phenomenon has led to the suggestion that companies should consider dropping the .ai top-level domain or refraining from adding 'AI' to their names unless they have substantive technology behind the label. These proposals mirror earlier debates around dot-com branding and emphasize the importance of substance over symbols.

## Future Directions

As AI technology continues to evolve, the line between genuine AI integration and superficial labeling may shift. Tools like [generative-ai](https://www.wikiprompt.org/wiki/generative-ai) have made it possible for even small companies to incorporate AI into their products through application programming interfaces, which could reduce the incentive for outright fabrication. However, the practice of overstating - for example, calling a simple API call a proprietary model - is likely to persist.

Clarity could be improved through standardized labeling or disclosure requirements. The European Union's AI Act and other regulatory frameworks may eventually require companies to disclose the use of AI in high-risk applications. These measures, combined with informed consumer behavior and investor due diligence, could help curb the most egregious forms of AI washing.

## See Also

- [ai-now-institute](https://www.wikiprompt.org/wiki/ai-now-institute)
- [artificial-intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence)
- [machine-learning](https://www.wikiprompt.org/wiki/machine-learning)
- [large-language-model](https://www.wikiprompt.org/wiki/large-language-model)
- [generative-ai](https://www.wikiprompt.org/wiki/generative-ai)
- [openai](https://www.wikiprompt.org/wiki/openai)
- [anthropic](https://www.wikiprompt.org/wiki/anthropic)
- [google-deepmind](https://www.wikiprompt.org/wiki/google-deepmind)

## External links

- [Wikipedia: AI washing](https://en.wikipedia.org/wiki/AI_washing)

---
Source: https://www.wikiprompt.org/wiki/ai-washing
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-11-09T12:23:13.616806+00:00