Most companies now have a policy on employee AI use, but only 17% of the risk professionals surveyed feel well prepared to manage AI risks.
Key Takeaways
- 46% of the risk professionals surveyed say their company trained employees to judge when to trust an AI tool’s answer and when to question it.
- 72% say their company has a policy on how employees use AI, up from 58% last year.
- 56% say their company is considering AI tools that can carry out tasks on their own. Of those companies, 58% haven’t assessed the risks.
- 17% of respondents feel very prepared for AI risks.
Only 46% of risk and compliance professionals say their company has trained employees to judge when to trust an AI tool’s answer and when to question it. The figure comes from the 2026 New Generation of Risk Report, a survey of 259 professionals published by Riskonnect, a risk management software company.
Most companies now have rules for how employees use AI, but fewer than half have taught employees to tell when its answers are wrong. More than half of companies are also considering AI tools that carry out tasks themselves instead of just answering questions, and most haven’t yet assessed the risks.
More companies have AI policies, but fewer say which tools are allowed
72% of respondents say their company has a policy on how employees use AI, up from 58% last year. Only 54% say their company has clear rules on which AI tools employees can use and how. Riskonnect said the policies are “stopping at the surface.”
Training on checking AI answers trails the policies
46% say employees have been trained to judge AI answers, compared with the 72% whose company has an AI policy. Fewer still have trained every employee on AI risks. 35% say their company has done this for the risks of generative AI, up from 17% in 2023.
Most companies considering AI that acts on its own haven’t assessed the risks
27% of respondents say their company already uses AI tools that carry out tasks on their own. Another 56% say their company is considering them.
Of the companies already using these tools, 84% assessed the risks. Of the companies still considering them, 58% have not. Across all respondents, only 18% say their company has trained or briefed every employee on the risks of AI agents.
Few respondents feel very prepared for AI risks
17% of respondents say they feel very prepared to handle the risks AI brings and to oversee their company’s AI use. 62% feel somewhat prepared.
23% say their company has a plan for AI threats such as deepfakes and fraud carried out with AI. 11% say their company has a budget set aside for reducing AI risks.
The report advises companies to teach employees when to trust AI, when to challenge it, and how to check its results. “Deploying a tool isn’t the same as preparing people to use it well,” it says.
The 259 respondents work in risk, compliance, or resilience at organizations worldwide. 36% are managers, 24% are directors, 24% are analysts, specialists, or coordinators, and 15% are vice presidents or C-level executives. 37% work at companies with more than 5,000 employees. The largest groups work in financial services and government.
