Executives are twice as likely as rank-and-file employees to say they trust artificial intelligence more than their own judgment. That is the conclusion reached by researchers at Charter, a company that studies the workplace. This reliance on AI could become a problem: top managers’ decisions affect not only individual tasks but entire companies, and a confident response from a neural network does not necessarily reflect reality, Axios reports.

The higher the position, the greater the trust in AI

According to Charter’s research, 32.8% of executives said they trust AI more than their own judgment. Among rank-and-file specialists, that figure was 16.3%, while among middle managers it was 34.8%.

The survey was conducted in August among 5,002 American office workers who use AI at work. All of them worked at companies with at least 100 employees. Participants included 2,952 rank-and-file employees, 1,547 managers, and 503 executives, including vice presidents, senior leadership members, business owners, and company founders.

This suggests that the people organizations expect to have experience, expertise, and the ability to make independent decisions are more willing than others to rely on algorithmic recommendations.

At the same time, the study measured self-reported trust in AI specifically. It does not show how often executives actually follow the advice of neural networks when making important decisions or what consequences that leads to.

When AI replaces independent thinking

Researchers use the term cognitive offloading — “cognitive offloading.” This refers to the use of external tools to reduce mental effort, such as delegating part of analytical work or decision preparation to AI.

According to report co-author Brian Elliott, a former Slack executive and now a Charter employee, top managers have particular reasons to rely on such tools. They are constantly expected to provide quick answers to a wide range of questions, and AI makes it possible to create an impression of competence even when there is not enough time for a full analysis.

Elliott notes that neural networks can reinforce some leaders’ habit of appearing confident in their correctness even when they do not have sufficient grounds for doing so.

AI is indeed good at quickly drafting summaries, synthesizing information, and suggesting plausible solutions. But a convincing formulation does not guarantee that the response takes into account the specific circumstances, constraints, and details of the task.

Why quick prototypes do not become finished products

Elliott cites the example of leaders at tech companies who use AI to create product prototypes over a weekend and then expect their teams to turn those mockups into working solutions just as quickly.

The problem is that a prototype and a finished product are not the same thing. Between them may lie complex engineering work, testing, bug fixing, and countless details that are not visible in an impressive demo.

AI handles preparing the big picture and brief managerial takeaways well, but it may be less useful when an exact answer to a specific technical problem is needed.

That is why excessive trust in neural networks can create a gap between leadership expectations and the team’s actual capabilities. If a manager takes AI’s confident answer as a verified solution, they risk assigning employees impossible tasks or making decisions without taking important circumstances into account.

Employees fear AI will make them unnecessary

Overall, Charter’s report gives a positive assessment of artificial intelligence’s potential in the workplace. Among surveyed AI users, 42.7% said it had improved both their professional and personal lives.

However, researchers also found another contradiction. In additional qualitative interviews with 91 employees, participants spoke about anxiety over the possible loss of their jobs. The better people master AI, the more clearly they understand how many tasks can be done with its help without the previous level of human labor.

A paradox emerges: higher productivity thanks to AI may simultaneously intensify fears that employees will become less in demand.

For executives, this threat looks different. According to Elliott, the likelihood that an AI CEO will one day appear and lead a company instead of a human is extremely low. As a result, top managers are largely spared the same sense of professional insecurity experienced by their subordinates.

But it is executives who make decisions with the broadest consequences. If they begin using AI not as a tool for testing hypotheses and analyzing data, but as a substitute for their own judgment, mistakes could affect the entire organization.

The key question is not whether leaders should use AI, but whether they retain the ability to evaluate its answers critically. The higher the cost of a decision, the more important it is to distinguish a persuasive algorithmic recommendation from a truly well-founded conclusion.