Getting ahead of the AI wave: Why you should implement an AI workplace policy

In recent years we have seen significant developments in the technological capability, and popularity, of artificial intelligence (AI) tools. The wave shows no signs of stopping!

AI tools can unlock opportunities for businesses to increase workplace efficiency. However, the potential misuse of such tools comes with several risks, including inaccurate or biased output being relied upon, as well as data protection issues and intellectual property concerns. These points could result in reputational damage, loss of business and even legal claims.

It is paramount that employers get ahead of this wave and implement workplace policies which deal with the use of AI tools by employees.

The purpose of such a policy would be to promote the benefits of AI tools whilst acknowledging the possible legal, ethical and financial implications of their misuse. Setting out clearly what is and isn’t permissible would make it easier for you to take any necessary action following an employee’s misuse of AI tools, including disciplinary action where appropriate.

Even you are not seeking to actively promote and invest in AI tools currently, having a policy is still recommended. Individual employees may already be using AI tools of their own volition, a concept being referred to as ‘shadow AI’, so a policy is key to make clear what they are and are not allowed to do. It is far too risky for any employer to allow unfettered use of AI tools in their workplace.

What should an employer’s AI policy include?

1. Which AI tools are permitted

Various AI tools exist, including ChatGPT, Microsoft Copilot and an array of industry specific tools. They vary in their approach to the data which is submitted to them, particularly regarding confidentiality. The recommended approach would be to identify specific authorised tools and to state that only such tools are permitted.

2. What activities the use of AI tools is permitted for

AI tools can assist with various tasks, including, drafting text, proofreading, summarising documents and completing research. We recommend that your policy specifically includes detail of any specific tasks that you believe would benefit from the use of AI and/ or any specific tasks for which the use of AI tools is prohibited.

3. An emphasis on human oversight

Whilst incredibly capable, AI tools do of course make mistakes. Your policy should emphasise that your employees ultimately remain accountable for the output produced by an AI tool that they choose to use. This is particularly important if any output is sent to a third-party (e.g.. a client).

4. Details of any types of information that should never be submitted to AI tools

Due to data protection concerns, including the requirements of the employer’s policies with both its employees and third parties (i.e. clients and suppliers), it is important to specifically prohibit the use of any confidential information being entered as a prompt unless you can be sure that the AI tool has appropriate safeguards for the data. It would also be necessary to consider how intellectual property rights owned by third parties would restrict the data that can be submitted (e.g. copyrighted work should not be input without the copyright owner’s consent).

5. Details on any record keeping and monitoring

The policy should make it clear that you reserve the right to lawfully monitor employees’ use of AI tools. This could be for individual performance or disciplinary related purposes, or to more broadly review the accuracy of approved AI tools, the frequency of their use within the workplace and their overall impact on business performance. Most AI tools should automatically keep a record of their prompts and outputs, but the policy should prohibit employees from deleting such records.

An AI policy is likely to interact with a number of other policies, including your data protection policy, disciplinary policy and IT and communication systems policy, so it will be important to cross-check them for consistency.

Please do get in touch with a member of our Employment Team if you require assistance in drafting an AI policy for your workplace.

Please note that this article is specifically about the use of AI tools by employees for business purposes. Specific considerations would arise if employers are intending to use AI for their own people management functions, such as for recruitment or performance monitoring purposes, particularly if automated decision-making is being considered. Our Employment Team can assist further with this.

Please note that this update is not intended to be exhaustive or be a substitute for legal advice. The application of the law in this area will often depend upon the specific facts and you are advised to seek specific advice on any given scenario.