Next Year Isn’t 2027—According to Google AI

To err is human. Every January, the same silly mistake appears. Many of us find ourselves filling out forms, signing emails or answering questions instinctively in reference to the previous year. Even weeks into a new one, people hesitate for that fraction of a second to scribble 2025 instead of 2026. In the imperfect world, that's totally acceptable.

But from an AI?

Surely not.

Time, dates, the calendar are supposed to be absolutely fixed. Unless you're in an upside down world where time moves backward or skips a few days, hours, years; we know the chronology of it all. But a confident assertion of the wrong year, delivered with mechanical certainty, feels… different and totally unacceptable.

But this is exactly what happened recently! When asked the world's most popular and sought-after search engine on the internet, Google, about what year comes after 2026, one would expect the answer to be 2027 no doubt about it. Any sane man who knows how to count would know. However, Google's AI believes it will be 2028.

You're not alone to be confused. According to Google's AI, "It is 2026 right now, and will be 2028 next year, followed by 2027". Google's AI overview reads that "2027 is two years from now as the current year (2026) followed by 2027and then 2028. Next year will be 2028, with 2027being the year after that."

Mindboggling as it sounds, the AI overview does say that. But why would that be a mistake or is it? Google's AI summaries are already infamous for hallucinations, yet you might assume a question as simple as the calendar would be beyond its grasp. Fat chance.

Screenshots circulating online confirm the claim that the AI overview on the web engine does repeatedly answer incorrectly. In another instance, the AI insists it’s currently 2025. Reddit threads have flagged the issue for more than a week. And it’s not just Google: OpenAI’s ChatGPT also falters, asserting, “No 2027 is not next year”—emphasis included. Clearly, something is misaligned. If you were marking the arrival of the singularity, it might be wise to push that date back a little.

The incidence does draw attention to the growing overreliance on artificial intelligence to help one manage the day. But also clearly states that they cannot be blindly considered advisors, assistants, arbiters of “fact.” The more people lean on AI for everything from weather forecasts to historical dates, the less they check, the less they pause. Small errors—like insisting next year is 2028 which honestly is a case of commonsense—might seem trivial. But each one chips away at the instinct to verify, to think critically. With a machine, doubt is absent.