'This is a training and trust problem, not an AI problem': Barely any workers are verifying information given to them by AI

A recent survey conducted by O’Reilly Media reveals that a significant number of UK employees are failing to verify information provided by AI before taking action. Only 24% of workers check the accuracy of AI-generated responses, while 39% have followed unsuitable or incorrect AI advice. O’Reilly’s research surveyed 2,000 UK workers and found that under a quarter (24%) of employees bother to confirm the results or directions from AI, despite AI chatbots being favored over colleagues and managers for work-related matters.

The survey highlights a critical gap in AI training for employees. Only 16% of respondents received AI training relevant to their role, and 35% reported receiving no AI guidance or training at all. This lack of training, combined with the rapid adoption of AI technologies, raises concerns about the reliability and safety of AI in the workplace. Alexia Pedersen, SVP International at O’Reilly, emphasizes that the issue is not with AI itself but rather with the training and trust surrounding its use.

The survey indicates that AI is being embraced as a primary source of information and advice by employees, but the necessary safety nets, such as verification habits and trusted expertise, have not been put in place. Pedersen stresses that the solution lies in closing the gap with comprehensive training and fostering a culture of verification. She argues that organizations must invest in role-specific AI training and promote a culture of skepticism and verification to turn AI adoption into a productivity advantage.