Artificial intelligence is helping many office workers produce work beyond their experience level, making it harder for employers to assess underlying skills.

A survey of 9,684 working adults across the US, UK, Canada, the EU and Latin America by Use.AI found that 52% believe AI has made them appear more experienced than they are.

Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram

AI Is Raising Output Faster Than Skills

Some 64% said they had used AI to complete work they could not have done independently, while 43% said it enabled them to take on responsibilities they did not feel qualified to handle.

Another 35% said they would struggle to perform parts of their current jobs without AI, and 25% worried their employers see them as more capable than they are. Meanwhile, 39% had submitted AI-assisted work without telling their manager, and 30% had accepted praise for output substantially produced by AI.

For 19% of respondents, AI-assisted work had contributed to a promotion.

Should Employees Disclose AI Use?

As AI becomes embedded in everyday software, requiring disclosure of every interaction may be impractical.

“I do not think companies should require employees to disclose every interaction with AI,” Ihor Herasymov, co-founder and chief executive of Use.AI, told Euronews. He said disclosure should apply when AI materially shapes the work.

“If it generated a significant part of an analysis, recommendation, presentation, code or other consequential output, employees should disclose that assistance,” Herasymov said. Employees should remain responsible for understanding, verifying and defending the work they submit.

Managers Need New Ways To Assess Performance

AI-assisted workers are not necessarily unqualified, but finished work now reveals less about the person who produced it.

“Finished output still matters, but it is becoming a less complete measure of capability,” Herasymov said. Managers should also assess whether employees can explain their reasoning, identify weaknesses in AI-generated answers and make sound decisions when the technology fails.

Problem framing is another key skill, he said: “Can someone define the right question, challenge an assumption and explain why one course of action is better than another?”

Organizations are still developing ways to distinguish what employees can produce with AI from what they actually understand.

AI Tool Or Dependency?

The finding that 35% of workers would struggle without AI raises questions about whether augmentation can become dependency.

“Yes, I think that finding deserves to be taken seriously,” Herasymov said, arguing that the risk emerges when workers cannot recognize incorrect AI output or make sound judgments when the system has no reliable answer.

AI can make workers faster and expand their capabilities, he said, but weaker independent judgment is a trade-off employers and technology companies need to address.

AI Autonomy Is Accelerating

The challenge is growing as AI systems become more autonomous. Ventureburn, citing METR data, reported that the time required for AI autonomy to double has fallen from an eight-month trend to 4.7 months.

Autonomous capabilities increased 1,400% year over year between early 2025 and early 2026, while AI tool downloads reportedly rose from 15,000 to 11.8 million. Publicly available MCP tools increased 35-fold to about 177,000.

MCP, or Model Context Protocol, lets AI assistants connect directly to applications and data sources to perform tasks. As AI takes on more work with less human intervention, employers may need to assess not only the final output but also the judgment behind it.