The use of AI agents is rapidly growing in the UK and the US as it cuts down on what is perceived as tedious tasks and improves productivity and efficiency.

Research from Pegasystems, however, shows that about a third of workers (33%) do not trust AI agents’ reliability or quality.

The research found that almost six in ten (58%) workers are using AI agents daily in the survey conducted by YouGov of over 2,100 workers.

The study defined AI agents as software programmes or tools powered by AI that can generate content, analyse data, automate repetitive processes, and provide recommendations,

The main benefit cited, by about four in ten workers, was the automation of repetitive, tedious tasks, as well as a reduction in the time spent searching for information related to work, and the summarisation of meetings.

Despite these major benefits, workers have yet to fully trust agentic AI, with a third (33%) saying they were worried about the quality of AI agents’ work.

A third (32%) also noted a lack of emotional intelligence and human intuition from the AI agents, an interesting critique for a computer programme.

Almost more concerningly, given the commonality of use, is that three in ten (30%) say they do not trust the accuracy of the AI agent’s responses.

Recommended reading

• Capgemini: How to Build a Responsible Generative AI Strategy
• MIT Tech Review: Responsible AI Principles Urgently Needed
• Report: 58% of Business Leaders Have No AI Training

Around four in ten (39%) report that they do not “feel good” about submitting AI-generated work, which could be a deeper sign of integrity, or a bigger sense of mistrust in the machines.

Pegasystems’ analysis of their data says that AI agents should be treated as a collaborative partner, and that agentic AI is able to continuously learn and therefore decrease errors. Issues remain about trust and accuracy in high-stakes environments, however.

Effectively training workers on how and when to use AI agents appropriately is not just a safety issue, but a trust issue. Empowering workers will ensure that agentic AI can safely integrate into system flows, and give workers the confidence to work with advancing AI systems.