It seems artificial intelligence has become the most confident person in the office. The funny thing is that according to new research, some bosses may trust it more than they trust themselves.
Now that is a slightly terrifying thought.
A new survey from workplace research company Charter found that 32.8% of executives trust AI more than their own judgement. Managers showed even greater confidence, with the figure rising to 34.8%. And among ordinary employees, that number fell to 16.3%. So it seems the people making the biggest decisions may also be the most willing to ask a chatbot what it thinks.
Axios reports that Charter surveyed 5,002 American office workers during August. Everyone who took part worked for a company with at least 100 employees and used AI in their job. The group included 2,952 ordinary workers, 1,547 managers and 503 executives. Now before we go any further, let’s make one thing clear. The survey measured what people said they trusted. It did not establish whether executives actually follow AI advice when making serious decisions.
But the results are still rather interesting.
Executives were 2.5 times more likely than workers to trust information produced by AI. They were also more than twice as likely to trust it over the opinion of a knowledgeable colleague. You know, an actual person who may have spent several years learning how the company works. I have to wonder how that colleague must be feeling after the survey.
Now it’s fair to assume that when executives use AI, they receive enormous amounts of information and must make decisions quickly. And we know that an AI tool can summarise reports, create presentations and produce an answer within seconds. No further meetings are required and everything is instantaneous. The problem comes in when speed starts looking like intelligence.
AI can produce something that sounds extremely convincing even when it has misunderstood the question, ignored important context or simply invented part of the answer. Unless the problem is picked up, it will not raise its hand and admit that it has no idea what is happening. Instead, it walks into the room with the confidence of someone who did not read the report but has volunteered to present it anyway.
Workers may understand this better because they often have to deal with the output. Charter’s research found that only 10.5% of AI users felt the technology had made their working lives much better. Less than 40% had received hands-on training, even though workers valued that training far more than company policies and written guidelines.
And that may be the problem. Companies can certainly tell employees to use AI, but if they do not explain where the technology works and where it could create trouble, people are left to figure that out by trial and error. There is also a difference between asking AI for help and handing over your judgement entirely. I use a calculator because I do not want to perform long division during lunch. That does not mean I would allow it to decide who should lose their job.
AI can be useful for finding patterns, producing a first draft or helping someone get through a pile of information. But a first draft is just that: a first draft. Somebody still needs to check it. A recommendation still needs context. And an important decision needs a person willing to accept responsibility for it and remain accountable. There is no shying away from that.
If a decision goes wrong, “the AI suggested it” cannot become the modern version of “the dog ate my homework”. A machine does not have to explain itself to employees, customers or a board. But you can bet that the human who trusted it does. Now in no way does the Charter survey prove that executives have stopped thinking for themselves. What it shows is how easily a useful tool can become an authority figure when it speaks quickly enough.
There is no problem with AI occupying a seat inside the meeting. But perhaps we should think carefully before giving it the deciding vote.
