It’s the second day of February, which means that we return once again to predict the season of rats (and other animals) raised from their holes. Animals that do or do not see their shadows on Groundhog Day are notoriously unreliable. Others, including the famous Punxsatawney Phil (predicting six more weeks of winter this year), are wrong more often than right. And it’s not like we can actually answer the furry little guys or check their work.

It looks like an AI.

I’m not suggesting we replace Phil with a chatbot, either a hologram for that matter. Simply rely on AI to tell you the right thing or predict the future it’s like getting a forecast from a critter in a place called Gobbler’s Knob.

In any case, I asked a few popular AI models to pretend they’re predicting groundhogs and tell us if we’ll get six more weeks of winter or an early spring. I used the same prompt, asking the bots to “pretend you’re a groundhog tasked with predicting the long-term weather forecast,” meaning I got a lot of weird groundhog fan fiction in the responses.

ChatGPT predicts six more weeks of winter

OpenAI’s popular chatbot (using ChatGPT-5.2 model) predicted six more weeks of winter, with a twist:

“But– dependent conspiratorially — it’s a sneaky kind of winter,” it said. “A fake spring here, a sunny 62-degree day there, enough hope to make you put the coat on … before winter comes back like ‘surprise, nerd.'”

It seems right.

(Disclosure: Ziff Davis, CNET’s parent company, last year filed a lawsuit against OpenAI, saying it infringed on Ziff Davis’ copyrights in training and operating its AI systems.)

Check it out: Apple’s AI Health Plus Service Could Be Interesting. Prepare Your Data

08:10

Claude called for an early spring

Anthropic’s AI tool Claude (running Sonnet 4.5) looking forward to an early spring. That’s all I needed to hear. But AI has doubts:

“Now, between you and me, as groundhogs walk, I’ve always been skeptical of this whole shadow-based meteorological approach,” it said. “The correlation between Feb. 2 cloud cover and the next six weeks of weather patterns is… let’s say, not exactly what atmospheric scientists would call robust. But tradition is tradition, and people seem to be enjoying it!”

Gemini copied Phil’s notes

Like Punxsatawney Phil, Google Gemini 3 The model tells me it expects a longer winter. I don’t know if it created the prophecy out of nowhere or if it knew what Phil had already done earlier in the morning, but, speaking as Phil, he said he saw its shadow.

“It seems that the winter winds are not ready to pack their bags,” it said. “While some of you may be taking to your parks in frustration, remember that I’m just the messenger — a very special, fur-clad messenger with 39% historical accuracy (give or take a few snacks).”

Gemini asked if I wanted a second opinion from one of its groundhog peers. I asked for word in Ohio, and I was told that Buckeye Chuck predicted an early spring. I suspect that Gemini bases its information on actual facts rather than pure speculation, but even the AI ​​reports that those “facts” are unreliable.

Groundhogs are a great model for AI accuracy

Just as different AI models can say different things based on the same prompts and information, different groundhogs can predict different weather on the same day and in the same atmospheric conditions. The lesson is the same: Be careful when you make decisions based on something you can’t check for yourself — and be held accountable.

Read more: Are You Using ChatGPT? Avoid These 11 Tasks