What is AI washing? | IT Pro
AI washing is becoming a widespread issue in the tech industry. This is the practice of companies liberally labelling their products or services as ‘AI-powered’, often in an attempt to capitalize on a technology’s hype, when in reality their use of generative AI or even AI more broadly is at best minimal, at worst non-existent.
Wishing to gain market attention by associating your offerings with a popular emerging technology is nothing new, but it does make it harder to distinguish products and services that are genuinely leveraging an emerging tech in new and exciting ways.
“This trend bears similarities to the misleading ‘greenwashing’ practices seen in the sustainability sphere,” notes Chris Carreiro, CTO of IT service provider Park Place Technologies.
“Just as companies might overstate their environmental efforts, AI washing involves exaggerating the role and capabilities of AI in a given solution. In simpler terms, it’s a marketing tactic that creates a deceptive impression of technological advancement.”
In recent years this ‘washing’ of capabilities has become particularly prevalent in AI – to date analyst firm Gartner has published over 70 reports that mention this phenomenon.
It’s almost like school playground peer pressure, where if you’re not talking about AI, you risk looking like you’re falling behind, according to Justin Sharrocks, managing director EU/UK at solutions provider Trusted Tech.
This has led to a swathe of tools marketed as AI-driven, but in reality, they’re just basic automation or rule-based systems, without the complexity of data actually being processed by an AI model.
AI washing – the risks
Organizations must ensure they do their due diligence when considering investing in a solution that pertains to be AI-powered, as they can be harmed in many ways if it isn’t as advanced as it claims.
Such investments could cost the company financially, while also failing to achieve the intended business objectives such as improved productivity and/or efficiency. There are also the security risks involved, especially in regulated industries where AI could help prevent fraud or compliance violations.
These can leave organizations vulnerable in areas they thought they were protected, opening them to brand or reputation damage due to data exposures or service failures, “resulting in fines, loss of customer trust and ultimately market share,” says Bern Elliot, distinguished VP analyst at Gartner.
“A supply chain concern is that some newer generative AI methods require special cybersecurity precautions, yet if a vendor isn’t specific about what their AI is, the buyer assumes there are the correct safeguards. But then that product becomes an element of another and the safeguards may be forgotten because the use of different types of AI was not clear,” he continues.
Elliot also urges leaders to consider risks such as prompt injection, which he says can be a serious risk. Without vendor transparency, he says, buyers can’t guarantee that any steps are being taken to mitigate this threat.
Finally, AI washing erodes trust in the technology’s potential. When the AI bubble around a product inevitably bursts due to this practice, it will create skepticism and reluctance towards legitimate AI solutions, says Sharrocks.
“This damaged trust hinders the adoption of AI technologies that could bring substantial benefits, ultimately slowing down progress and innovation across the business landscape.”
How to spot AI washing
Identifying AI washing amidst the current hype can be a challenge, especially given the lack of universally accepted definitions and clear regulations surrounding AI. This ambiguity creates fertile ground for companies to exploit the AI label without necessarily delivering on its promises.
To navigate this landscape, businesses need to be vigilant and ask critical questions, says Carreiro.
Firstly, businesses must prioritize substance over buzzwords, he says:
“Look beyond marketing claims and delve into the specifics of how AI is implemented within a solution. Is it a core component driving key functionalities and decision-making processes, or is it merely a sprinkle of automation or enhanced analytics rebranded as ‘AI lite’? Carefully explore whether organizations are truly implementing AI into their processes and businesses, or whether this is simply a repackaging of existing analytics and machine learning techniques.”
Secondly, he advises that you request concrete evidence of the AI’s impact.
“Ask for case studies, performance metrics and quantifiable improvements achieved through the solution. A genuine AI implementation should yield tangible results and demonstrable value beyond what traditional analytics or automation can provide. If a vendor struggles to provide clear evidence of the AI’s impact, it’s a red flag.”
The regulatory domain
While it may feel a bit like the Wild West right now, businesses aren’t being left to navigate this new landscape all alone. While the going may have been slower than we’d like, governments and regulators are starting to crack down on misleading AI claims.
Companies caught AI washing may face fines, legal action and reputational damage. For example, investment firms Delphia and Global Predictions were fined by the US Securities and Exchange Commission (SEC) last year for making misleading claims about their use of AI.
Furthermore, regulation around AI will continue to grow, meaning businesses will have to adhere to stricter standards and be held more accountable for false claims.
“This is especially true with the recent actions of regulatory bodies in the UK and EU, which have started to take a closer look at misleading AI claims,” says Sharrocks. “As AI continues to evolve, regulators are likely to introduce clearer guidelines around AI marketing. Just like we’ve seen with data privacy laws like GDPR, AI-specific regulations will likely emerge to ensure that businesses market AI responsibly and accurately. This will help prevent companies from engaging in AI washing and create more transparency for customers,” he concludes.
Source link
