By C. Perera, JadeTimes News
What is 'AI Washing' and Why is it a Problem?
Amazon faced scrutiny this year when reports surfaced questioning the "Just Walk Out" technology in its grocery stores. This AI driven system, used in Amazon Fresh and Amazon Go shops, allows registered customers to pick their items and leave, with AI handling billing via facial recognition, sensors, and cameras. However, reports in April indicated that instead of solely relying on AI, around 1,000 workers in India manually checked almost three quarters of the transactions.
Amazon quickly dismissed these reports as "erroneous," stating that the Indian staff were reviewing the system rather than footage from all stores, likening it to human reviewers common in any high accuracy AI system. Despite this, Amazon announced a reduction in the number of stores using Just Walk Out.
This situation exemplifies a growing concern about companies making exaggerated claims about their AI usage, a practice termed "AI washing," akin to environmental "greenwashing." AI refers to computers learning and solving problems after training on large datasets, with generative AI, such as chatbots like ChatGPT, Google’s Gemini, and Microsoft’s Copilot, gaining significant attention for creating new content.
AI washing manifests in several ways companies might claim to use AI when they employ basic computing, overstate AI's effectiveness, or suggest their AI is fully operational when it isn't. Some firms merely add AI chatbots to non AI software.
OpenOcean, a UK and Finland based tech investment fund, reported that in 2022, 10% of tech start ups mentioned AI in their pitches, which rose to over a quarter in 2023 and is expected to exceed a third this year. Competition for funding and the desire to appear cutting edge drive some companies to overstate their AI capabilities, as noted by OpenOcean team member Sri Ayangar.
MMC Ventures' 2019 study revealed that 40% of "AI start ups" used minimal AI, a problem that persists. Simon Menashy of MMC Ventures notes that cutting edge AI is now widely accessible, leading many firms to add chatbot interfaces to non AI products. Dougal Dick of KPMG points out that the lack of a single AI definition enables AI washing, causing businesses to overpay for technology and miss operational goals, and making it harder for investors to identify truly innovative companies.
Regulators are beginning to address AI washing. The US SEC charged two investment advisory firms with false AI claims earlier this year, indicating stricter enforcement ahead. In the UK, the Advertising Standards Authority already has regulations against misleading AI claims. Michael Cordeaux from Walker Morris highlights ASA investigations into exaggerated AI advertisements.
Sandra Wachter of Oxford University warns against the overhyped use of AI, citing examples like AI powered electric toothbrushes. She also emphasizes AI's environmental impact, noting its significant contribution to climate change. Wachter calls for a more measured discussion about where AI is genuinely beneficial.
Advika Jalan of MMC Ventures believes AI washing might decline as AI becomes ubiquitous, making "AI powered" a less significant differentiator, akin to stating a company is "on the internet."
Which kind of images add to this