Do you need an AI policy?

We are inundated with Artificial Intelligence (AI) tools. Examples include image generators such as Dall-E, Stable Diffusion, and Midjourney, and text generators such as ChatGPT, and Jasper. Google and Bing search tools also use AI to generate search results.

AI has been used for several years for things like recommendation engines, but only recently have AI tools been generally available for use by everyone.

It is inevitable that we will use AI tools in some fashion. But we need to keep in mind that AI carries several legal and ethical issues.

Depending on the entity, the nature of the AI tool, and how the output will be used, relying on them might cause issues.

Businesses and organizations should consider whether they should implement rules around the use of AI tools in an AI policy, or an addition to their existing technology use policy. Those rules might prevent certain activities, or might put a vetting process in place before use.

The goal could be to set a framework for ethical AI use. Also to control inappropriate rogue use of AI tools by employees who may not understand the risks and issues.

Some of these issues depend on how AI is being used. Using ChatGPT directly, for example, has more issues than if it is being used as a conversational interface layer to some other service you are using. It is important to understand that distinction and what the AI is actually doing in each situation.

AI policy issues to address

• Images and text created by AI tools are probably not protectable by copyright. So you can’t stop others from copying and using what you ask AI to create. If you don’t want others to use what you create, that could be an issue.
• AI tools generate output based on a database of existing images or text. Some of them retain the queries you make and any document you enter.
• Inputting anything that is confidential, sensitive, or personal should be avoided.
• The output from AI tools is notorious for being incorrect. It also parrots the biases contained in the material in its database. If AI tools are to be used to make decisions, the tool should be vetted for accuracy and bias. AI output should be carefully reviewed before it is relied upon or published.
• If the business wants to formally use AI tools, it should vet the tools to make sure they can be used in a legal way and comply with published ethical standards.
• Algorithmic transparency is desirable, and will no doubt be required at some point legally. Consider how transparent you should be with the public and your customers about how your AI makes decisions.
• Using AI to get ideas or inspiration for images, articles, or other creative works may not be something that needs to be disclosed. But from an ethical perspective, disclosure may be desirable if it is being used to make decisions or to create published material. To some extent, it is no different than passing off someone else’s work as your own. See, for example, how Wired recently described how it will and won’t use generative AT technology.
• Someone in the business should be accountable for the use of AI – similar to a privacy officer or CIO. That person would be the gatekeeper for AI issues and policies, vetting AI tools, and educating staff.

The last thing a business or organization needs is for an employee to use an AI tool in a way that embarrasses the business. Best to set the tone up front to prevent that from happening.

David Canton is a business lawyer and trademark agent at Harrison Pensa with a practice focusing on technology, privacy law, technology companies and intellectual property. Connect with David on LinkedIn and Twitter.

Image credit: ©putilov_denis – stock.adobe.com