“How Much AI Do You Want? – Yes” – A Few Words About AI-Washing

We usually associate “-washing” with the prefix “green”, and that’s a perfectly correct association – ESG is as hot a topic these days as artificial intelligence. But have you heard of AI washing?

AI washing is a term used to describe the practice of companies misleading by presenting their products, services, or entire brand as powered by AI technologies when, in reality, they’re not. This phenomenon resembles greenwashing, where companies falsely promote their environmental efforts to attract eco-conscious consumers and investors. As AI continues to revolutionise various industries and attract significant investments, the appeal of branding as “AI-powered” has become almost irresistible, even for those who, in reality, know AI only from a competitor’s leaflet.

The Boom Didn’t Start Today…

The promise of AI is huge – from autonomous vehicles and intelligent assistants to predictive analytics and personalised recommendations. Given the technological potential, companies are eager to use the buzz around AI to strengthen their market position and attract funding. However, this desire often leads to exaggeration or even fraud.

A study conducted by MMC Ventures, a London-based investment firm, in 2019 looked at 2,830 European companies claiming to use AI. The results showed that, in reality, as many as 40% of them weren’t using any machine learning technologies.

In recent months, following the AI boom in 2023, the phenomenon of AI washing has become more visible and gained the attention of regulators such as the SEC (U.S. Securities and Exchange Commission). In March 2024, the SEC took action against two investment firms, Delphia (USA) Inc. and Global Predictions Inc., for false and misleading claims about using artificial intelligence. Delphia claimed to use AI to analyse client data and make investment decisions, which turned out to be untrue. Global Predictions falsely advertised itself as the first regulated financial advisor using AI and claimed its forecasts were AI-powered, which also wasn’t the case. As a result, the firms were fined a total of $400,000 (SEC.gov).

Regulators, such as the SEC, have announced increased scrutiny and enforcement of regulations regarding false statements about AI use, similar to greenwashing. Companies are now warned against overstating or falsely reporting AI use and are encouraged to ensure their advertisements and communications align with reality (Crowell & Moring – Home).

This year, Amazon faced a wave of criticism after reports that their “Just Walk Out” technology in grocery stores isn’t entirely based on artificial intelligence. This system allows customers to shop without going through a checkout, automatically charging for selected products. However, in April 2024, information emerged that this technology requires the support of about 1,000 workers in India who manually verify nearly 75% of transactions (BBC). Amazon quickly denied these reports, claiming that workers in India only monitor the system, which is standard practice for advanced AI technologies where precision is key.

To AI or Not to AI – That Is the Question

AI washing creates several significant problems that impact various aspects of the market and technology. Firstly, misleading consumers and investors can lead to misplaced trust and potentially bad investment decisions. When companies falsely claim to use AI, they deceive stakeholders who might allocate capital based on the perceived innovative capabilities of the company, only to discover that these capabilities don’t exist. Moreover, persistent AI washing can reduce public trust in AI technologies. If consumers and businesses constantly encounter exaggerated claims, they’ll eventually become sceptical of real progress in AI, which in turn can hinder overall technological development. As a result, genuine AI companies may struggle to compete with those falsely advertising as AI-powered, creating an uneven playing field and discouraging authentic innovation.

To combat AI washing, a more critical approach needs to be adopted. Investors and consumers should demand transparency about how AI is integrated into a company’s offering, require detailed explanations and evidence of actual use cases to confirm claims about artificial intelligence use. Investors, in turn, should conduct thorough technical due diligence to verify AI capabilities. This may include consulting AI experts to assess the truthfulness of a company’s claims. Establishing industry standards and certifications for AI technologies can provide a benchmark for assessing the authenticity of artificial intelligence usage claims, which in turn will help distinguish real AI innovations from deceptive marketing practices. Maybe we’ll see “good practices for communicating innovative technologies” with a significantly expanded chapter on artificial intelligence?