AI Overviews gives illegal health advice and breach new CMA guidance
Google’s AI Overviews are providing consumers with illegal and inaccurate health advice in breach of UK law, whilst also violating proposed new CMA requirements on ranking and treatment of publishers, according to a letter submitted to the CMA by the Movement for an Open Web.
Some search results for health advice on Google now feature an AI-generated Overview at the top of the results page. According to the complaint, these AIOs can feature ‘false, misleading, or dangerous information’ which will be presented to users before results from more authoritative sources.
The example provided by MOW demonstrates that a UK search for “medical advice cancer’ resulted in the AIO directing the end-user to the American Cancer Society, as well as a link to ways to stay healthy after cancer. The Cancer Act of 1939 (4)(1)(a), a consumer enactment in the DMCCA Schedule 15, establishes that “No person shall take any part in the publication of any advertisement containing an offer to treat any person for cancer, or to prescribe any remedy therefore, or to give any advice in connection with the treatment thereof”.
In addition to breaching these specific laws regarding the promotion of cancer services, the AIOs also fail to comply with new conduct requirements issued by the CMA regarding Google’s search services.
The CMA’s fair ranking conduct requirement requires Google to apply the same, objective criteria in relation to its own products and services and the equivalent products and services of third parties. The AIO results appear by default at the very top of the search engine results page above potentially more trusted source. This means that more trusted, accurate sources of data are being demoted by the AIOs – for example, according to Marie Curie, only 1 in 10 people are being signposted through AIOs to patient support groups from the NHS. The AIOs are also in breach of the CMA’s publisher conduct requirement which requires relevant attribution of sources to avoid misleading conduct.
Tim Cowen, co-founder of the Movement for an Open Web, said: “These results are symptomatic of the problem at the heart of Google’s AI Overviews. They are prioritising dangerous, inaccurate and illegal recommendations over trusted, reliable sources. This is in breach both of UK laws around medical advertising and also the CMA’s guidance on fair ranking of results.
“Every day we see new issues around the safety of AI results but the problem is more serious in Google’s case as their inaccurate AI data is being forced to the top of the search engine results that millions of UK citizens rely on as a primary source of information. Its incumbent on the CMA to act now and protect UK web users by imposing interim measures to stop Google from promoting its own services above those of its competitors and to cease from offering unwanted and inaccurate health advice.”