Google AI Overview: AI is causing havoc in searches 🤖

Cliquez ici pour lire en français

Recently, Google launched the AI Overview feature, a new functionality aimed at improving search results using artificial intelligence. While this technology promises to simplify complex searches, it has also generated surprisingly erroneous advice. But the tech giant assures that it has learned from these chaotic beginnings and implemented improvements to address the main identified issues.

Dangerous Search Results ⚠️

Praised as the future of online search, Google’s generative AI, deployed via the AI Overviews tool, quickly showed its limits with a series of erroneous and outlandish answers that made the rounds on social networks. Among the blunders that went viral, the most decried was this absurd advice on how to make cheese stick to homemade pizza: add 25 ml of non-toxic glue to the sauce, a suggestion that came from an old Reddit thread where an Internet user was joking.

Beyond jokes, the AI also drew from discussion forums to provide unreliable, and potentially even dangerous answers according to certain screenshots disputed by Google.

Google Will Take Action 🚧

In a press release, Google explained that AI Overviews are designed to handle complex questions that would require multiple searches. Unlike classic chatbots, these overviews rely on integrated web ranking systems to provide relevant, high-quality links. However, the company acknowledges that some responses were erroneous and has taken steps to improve its algorithms.

Google announced several technical adjustments to avoid errors. Among the measures taken, the company has improved the detection of nonsensical queries and limited the inclusion of humorous and user-generated content. Additionally, trigger restrictions have been put in place for topics where AI Overviews have proven less useful, particularly in the areas of news and health.

AI hallucinations: a widespread Issue 🚨

Google is not alone in this situation. OpenAI, Meta, and other tech companies are facing similar errors in their AI systems. However, Google is particularly in the spotlight due to the wide diffusion of its technologies.

Google’s AI Overview feature showcases the potential of AI to revolutionize online search, but it also underscores the importance of vigilance against errors. Users must remain critical of AI recommendations, and companies must take responsibility for the accuracy of their products. As AI continues to improve, it is crucial to maintain a balance between innovation and caution to avoid advice as absurd as adding glue to a pizza.

While these bugs and hallucinated responses from Google’s AI may cool the enthusiasm of technophiles dreaming of a superintelligence, they above all remind us that this promising technology is still in its infancy. A warning not to be neglected, as tech giants accelerate headlong into the race for generative AI.

Do you often use generative AIs? Have you ever encountered such hallucinations? Let us know in the comments.