Google’s AI gets the year wrong and Elon Musk can’t resist commenting

 

It started with a question so simple it should have been impossible to get wrong: “Is 2027 next year?” Google’s new AI Overview feature somehow managed to turn that into a small internet moment.

Instead of giving a straight answer, the AI-generated summary told users that “2027 isn’t next year,” adding that “2026 is next year and 2027 will be the year after that”, while also, confusingly, presenting 2026 as the current year in the same response. Screenshots of the answer began circulating online, and it did not take long for people to start poking fun at it.

 

On X, one user joked that “2026 isn’t working great for Google,” a line that quickly became part of the pile-on.

Elon Musk soon joined in. Replying to the screenshot, the Tesla chief and founder of AI start-up xAI offered a dry verdict: “Room for improvement.” It was a brief comment, but a pointed one, and exactly the kind of jab that tends to travel far on social media.

 

Google’s AI Overviews are meant to give users quick, helpful summaries at the top of search results by pulling together information from sources like Wikipedia and Time and Date. In this case, the underlying facts were not wrong. The system simply tied itself in knots trying to explain them, and ended up producing an answer that contradicted itself.

 

It is the kind of mistake that feels funny because of how basic the question was. But it also touches on a broader concern that has followed the rollout of AI-generated search tools: these systems can sound confident and authoritative even when the answer is clumsy, misleading or just plain wrong.

This is not the first time Google’s AI summaries, or similar features from other companies, have been called out for strange or inaccurate responses. Past examples have ranged from harmless oddities to advice that was plainly unhelpful or incorrect.

 

In this case, the stakes were low. No one is going to miss a flight or lose money because a search result fumbled a calendar explanation. Still, it is a neat illustration of the limits of today’s AI systems, and of why users probably should not stop applying common sense anytime soon.

And, as Musk’s quick comment showed, there will always be someone ready to point it out when the machines get even the simple stuff wrong.