Survey finds workers worry AI dependence is weakening core skills

Half of employees say they depend too heavily on artificial intelligence at work, according to research by GoTo and Workplace Intelligence, a finding that points to a growing management challenge as AI tools become embedded in daily workflows. The survey of 2,500 employees and IT leaders also found that 39% of employees believe AI is making them less intelligent, 41% think it could damage their long-term career prospects and 30% say they can no longer function at work without it.

Dan Schawbel, managing partner of Workplace Intelligence and a future-of-work researcher, said the results show that workers are reaching a point where the benefits of AI adoption need to be balanced against the risk of skill erosion. He said resisting AI is not a practical answer, but neither is allowing the technology to take over decisions that require human judgment.

Schawbel’s view is that employees who retain value as AI systems improve will be those who use the tools to support their work, rather than replace their own thinking. He identified three habits that can help workers keep that balance.

Review where AI is being used

GoTo and Workplace Intelligence found that 60% of workers feel pressure to use AI to increase productivity. Schawbel said that pressure can encourage employees to apply the technology even when a task may not need it.

He recommends that workers keep a weekly record of assignments they pass to AI. At the end of the week, he said, employees should assess whether the tool improved the result and whether they could have completed the work themselves.

The purpose of that review is not to remove AI from the workflow. It is to separate uses that save time or improve output from uses that replace analysis, writing, judgment or problem-solving that the employee still needs to practice.

Attempt difficult work before asking AI

The survey found that 70% of employees acknowledged using AI improperly for sensitive or high-stakes work that calls for emotional intelligence or sound judgment. Examples cited by Schawbel included difficult conversations and decisions with material consequences.

Schawbel said employees should spend at least 15 minutes working through demanding tasks before turning to an AI system. That could mean drafting a first version, laying out an argument or reaching an initial conclusion before asking AI to test, refine or critique the work.

The distinction matters because AI tools can produce confident but flawed responses. In the same survey, 43% of employees said they had used AI-generated material even when they suspected it contained errors or fabricated information. Another 31% said they felt informal pressure to trust AI and stay silent about mistakes, while 14% said they reported AI errors to a manager and were told not to speak up.

Schawbel said workers and leaders will need to retain the ability to identify when an AI response is wrong, incomplete or unsuitable for the situation. That ability depends on regular practice with the underlying work.

Invest time in human skills

Employees surveyed by GoTo and Workplace Intelligence said creative thinking, emotional intelligence and judgment will be among the most important capabilities in an AI-driven workplace. Schawbel said workers should use time saved by AI to develop those skills rather than allowing them to weaken.

He suggested choosing one area to improve each quarter. Client-facing employees, for example, can seek situations that require them to read a room or address disagreement without relying on a script. Employees in analytical roles can practice forming and defending their own conclusions before checking a chatbot or outside commentary.

Schawbel said the workers most likely to remain useful as AI advances will be those who can combine machine assistance with independent judgment, rather than treating AI output as a final answer.

This story draws on original reporting from CNBC.