Policy and training needed to counter employee DIY AI risk

Governance, clear policies and training around ethical AI are urgently needed to mitigate against unauthorised usage.

Many employees are using generative AI tools at work without formal approval or guidance from their employers. More than a quarter (28%) of employees surveyed said they currently use generative AI at work, and over half do so without formal approval.

These are the headlines from a Salesforce survey, in partnership with YouGov in October which polled more than 14,000 employees in 14 countries including the US, United Kingdom, Canada, France and Germany.

I start every training session about AI in management and public relations with a table of the data management policies of large language models, including Bard, ChatGPT and Claude. Please do not cut and paste or upload documents to any web-based tool without checking its data management policies and asking yourself whether you trust the vendor.

The data management and security issues are clear, but there is also a growing ethical issue.

Salesforce reports that employees are engaging in questionable activities with generative AI, such as passing off AI-generated work as their own or exaggerating their AI skills.

64% have passed off generative AI work as their own and 41% would consider overstating their skills to get a job.

The Salesforce survey investigated the use of generative AI in the workplace, policies around generative AI, training on ethical/safe use of generative AI and attitudes towards generative AI's impact on careers.

The good news is that employees using unauthorised generative AI tools recognise this is a problem. Those using the tools admit that ethical and safe use of generative AI means adopting company-approved programs.

Most employees have not received training on the ethical use of generative AI. Nearly 70% of those surveyed globally said they have never completed or received such training.

Companies need clearer policies around the use of generative AI. 39% of employees said their employer doesn’t have a strong opinion on use of generative AI at work. Specific industries, such as healthcare, lag even further behind. This is a significant concern as it is a highly regulated market.

The solution is for employers to implement clear ethical guidelines and training on generative AI so employees can harness benefits while understanding and mitigating risks.

Thanks to Andrew Bruce Smith for highlighting the study.