AI helpers are writing computer code. Should we trust them?

Picture a artificial intelligence so small you could hold it on your fingertip.¹ Now imagine it doing something no one thought possible. That is what a team of researchers reported this month, and the finding has other scientists talking.

Not every test worked the first time.² “We broke a lot of equipment,” one researcher admits with a laugh. “Science is mostly about trying again.”

The researchers began by collecting data over several seasons.³ They recorded measurements every few hours, day and night. “We wanted to be sure we weren’t just seeing a one-time fluke,” explains one member of the team.

Students can get involved, too.⁴ Many research projects now invite volunteers to count, photograph or record what they see near their homes and schools. Those observations add up.

What comes next

At first, the results looked confusing.² Some measurements went up while others went down. But when the team plotted everything on a graph, a clear pattern jumped out.

Still, there are limits to what a single study can show.⁵ The researchers looked at one place during one period of time. Results elsewhere could be different.

This kind of work takes patience.⁶ Some samples had to be checked under a microscope, one at a time. Others were analyzed by computer programs that can spot tiny differences people might miss.

The researchers shared their results in a scientific journal this summer.¹ Other teams can now repeat the experiments to see if they get the same answers. That checking is an important part of how science works.

What the data show

The team compared its findings with records going back more than 50 years. Older data came from notebooks, museum collections and even old photographs. Putting it all together was a bit like assembling a giant jigsaw puzzle.

The findings could help in unexpected ways. Engineers, doctors and farmers all rely on this kind of basic research to design better tools and make smarter choices.

“We expected one answer and got a completely different one. That’s the best feeling in science.”

A member of the research team

Other experts who were not involved in the study say the work is careful and convincing. One calls it “a really clever way to answer a question that has bugged people for years.”

To check their idea, the scientists set up a careful experiment. One group was kept in normal conditions. A second group faced small changes in temperature, light or food. Everything else stayed the same, so any differences could be traced back to those changes.

Questions that remain

Why does this matter? Understanding how one small part of nature works often helps explain much bigger systems, from forests to oceans to the human body.

The team is already planning its next experiment. “Every answer leads to three new questions,” one scientist says. “That’s the fun part.”