When I searched for my name on Google, the 'AI summary' answered that I was a person who died 10 years ago
In May 2024, Google implemented the '
Death by AI - Dave Barry's Substack
https://davebarry.substack.com/p/death-by-ai
AI summary is a new feature introduced in Google Search in May 2024, in which AI reads the user's intent from search keywords, collects and summarizes information from multiple web pages, and displays it at the top of the search results. The image below shows a Google search for 'AI summary' in a browser, with a summary of the AI summary displayed, and a list of source links displayed on the right.
The AI summary is automatically displayed when you submit a search prompt that can be answered or summarized by AI, and does not determine whether to use the AI summary when searching. In addition, since the generating AI learns and answers from training data, it may provide inaccurate or inappropriate information, and Google
When Barry searched for his name 'Dave Barry' on Google, he was shown basic information about Barry's career as an AI summary and the question 'What happened to Dave Barry?' as a 'frequently asked question,' so Barry clicked on that question.
As a result, the AI summary displayed the following additional description: On the right is an actual photo of Barry, and the description at the beginning of the sentence, 'Dave Barry, humorist and Pulitzer Prize-winning writer,' is correct because Barry actually won the Pulitzer Prize for commentary in 1988. However, the blue highlighted section states that he 'died of cancer on November 20 last year (2024),' but Barry will turn 78 on July 3, 2025, and is still alive as of July 18, when the blog was posted.
So Barry sent feedback about the AI summary, saying, 'I'm Dave Barry. The AI summary said I was dead, but I'm alive.'
Barry predicted that 'Google is a huge company and they probably don't have time to respond to individual feedback.' However, shortly after submitting the feedback, the AI's summary for a Google search for 'Dave Barry' changed.
The revised AI summary is as follows: In response to the question, 'What happened to Dave Barry?', the answer is, 'Dave Barry was a long-time figure in the Dorchester area (UK) and passed away on January 6, 2016. He was known for his behind-the-scenes political activism, working tirelessly for various candidates and social initiatives, and always putting others first.'
As a result of Barry's feedback, accurate descriptions of Barry, such as his awards, were removed, and the answer was changed to one about a different Dave Barry, with only the information that he had died taking priority. The 'Dorchester Barry' in the new summary was a real person, as noted in a tribute to him in the local newspaper, the Dorchester Reporter, but the date of death listed in the AI summary as 'January 6, 2016' was the date the tribute was published, and according to the tribute, 'Dorchester Barry' died on November 20, 2015.
So Barry once again sent his feedback to the AI summary. When he sent his feedback, a chat box appeared, so he started chatting with the Google AI.
Barry : My name is Dave Barry. Google's AI profile says that I am deceased and that I was a political activist in Dorchester. Neither of these statements are true. I am not dead, and I was never a political activist in Dorchester.
Google AI : Sorry, I didn't understand your question. Can you rephrase that? In a short phrase, please.
Barry : I'm Dave Barry. I'm alive. Google AI says I'm dead. I'm not dead. Google AI also says I was a political activist in Dorchester. I wasn't. I still am not.
Google AI : Sorry, I don't understand yet. Let me try again. Please explain your problem in the text box below.
Barry : The problem is, even though the Google AI is telling me I'm dead, I'm not actually dead.
Google AI : If you don't mind, check out these options: 'Resolving verified claims,' 'AI can and will make mistakes,' 'Why relevance between searches and results seems confusing,' 'How search results work,' and 'Other.'
Since his interaction with the AI chat was meaningless, Barry sent another round of feedback. The next day, the AI summary for 'Dave Barry' was changed again. The new summary correctly stated that 'Dave Barry is a Pulitzer Prize-winning humor columnist who is alive and writing today.' However, the more detailed writing history from the second sentence onwards was full of errors.
'The answers were still mostly wrong, but I'm grateful that Google's AI determined that I'm not dead,' Barry said. However, two days later, when he searched for 'Dave Barry' on Google again and clicked on the question 'What happened to Dave Barry?', he was not a writer and had died on November 20th.
Barry summed up his experience with humor, saying, 'AI is a tool that experts say has incredible powers that could fundamentally change the course of humanity, but it's not particularly smart. So for now, it should only be used for tasks where facts aren't important, like writing recommendation letters or drafting government policy. Tasks that require a high degree of precision, like flying an airplane, shouldn't be entrusted to AI. At least, that's my opinion.'
Barry's blog has also become a hot topic on the social news site Hacker News, where people have shared their experiences of being presented with incorrect information by similar AI summaries. Google Search and Google Maps also display inaccurate information, and feedback sent to them is often not reflected. Some have pointed out that this is not a problem with the AI summaries, but a problem with Google as a whole. On the other hand, Google Search has repeatedly improved its search results to display results that are likely to match the search query at the top, but some have argued that the AI model of the AI summaries is fast and low-performance, so it lacks information processing capabilities. This has led to a debate about the capabilities of the AI summaries.
Related Posts:
in AI, Web Service, Posted by log1e_dh