AI-generated search summaries have surfaced sensitive suicide details at the top of Google results. Mental health organizations say this violates reporting guidelines. The issue raises new questions about editorial standards for AI content.
Media and publishing professionals are raising concerns after Google’s AI Overviews surfaced explicit details about a recent suicide at the top of search results, highlighting the risks of automated editorial decisions for sensitive topics. The incident has prompted renewed scrutiny of how AI-generated content handles reporting standards, especially when it comes to mental health and public safety.
This week, searches for “Caroline Flack brother” on Google prominently displayed the method of suicide in the AI Overview, placing it above publisher links and traditional news stories. Paul Flack, brother of the former Love Island presenter Caroline Flack, died at age 55 after being found unresponsive at his home, six years after his sister’s death.
The AI-generated summary included not only confirmation of his death but also the specific method, location, and additional details about the circumstances. This approach runs counter to guidelines from mental health organizations such as Samaritans and Mind, as well as the UK Editors’ Code, which advise against reporting suicide methods to prevent potential harm.
Nicola Agius, Reach’s SEO and Discover director, first noticed the issue after searching for news about Paul Flack’s death. She reported that Google’s AI Overview broke down the incident into sections, including subheadings like “Discovery,” and provided explicit details that are typically avoided in responsible news coverage.
Press Gazette was able to replicate the AI Overview on Thursday, finding that the method of suicide appeared in the first two paragraphs of the summary and again in expanded content after clicking “show more.” While the information was sourced from publisher reporting, it was made far more prominent in the AI summary than in most news articles.
Some publishers have since removed these details from their stories. For example, The Times, which was cited by Google for the suicide method, no longer included those specifics by Thursday morning. However, the Daily Star continued to list the provisional cause of death in a subheading, and Metro included it in the third paragraph of its article.
Mental health organizations have long advised against reporting suicide methods. Samaritans’ media guidelines urge publishers to avoid describing methods, especially in headlines, while Mind warns that mentioning specific methods or locations can be triggering and potentially harmful.
Lois Sparkes, head of safeguarding at Mind, described the situation as “incredibly worrying” and not in line with established reporting guidance. She noted that AI summaries can be especially triggering when they lack clear signposting to mental health support and present sensitive information as definitive fact. Sparkes called for AI-generated news content to be held to the same standards as traditional media and for companies to dedicate more resources to preventing the spread of harmful information.
Google has been asked for comment on the issue. The debate over AI editorial standards comes as publishers and platforms increasingly rely on automated tools for content distribution and search visibility. In a related development, Google has also been testing new AI features for its ad operations, as reported in a recent story about its Ask Ad Manager chatbot designed to streamline troubleshooting for publishers.
