The growing use of artificial intelligence in advertising comes with a critical question for marketers and agencies: How can they take advantage of this new technology without taking on too much risk? Here are six strategies:

Decide whether AI is right for your company

This is especially true for advertising agencies that pride themselves on creativity. How will clients react to the use of AI to generate ideas? Will they want to pay less? Will they expect the agency to take on all associated risks?

Agencies should also be aware that their current agreements with clients may prohibit the use of open-source software, which likely includes many AI platforms. Such agreements might also require the agency to create and transfer “ownable” content. This might not be possible with generative AI given the U.S. Copyright Office’s position that content that isn’t the result of an author’s “own mental conception” likely isn’t entitled to copyright protection.

Brands should consider whether they want AI platforms to ingest their data and use it for their own future purposes—which is generally how such platforms’ terms are structured. This means a brand might not be able to stop others from infringing on creative materials produced via AI and that such content may be used by the platform and its users in the future.

Determine whether AI output will be public or internal

Generative AI can create public-facing advertising or entertainment content such as concepts, text, images, video and music. AI software can also be used for internal analytical purposes. These uses include media buying and programmatic advertising, customer analysis and others that provide insight into the marketing process.

Generative AI currently presents the most risk to a brand or agency. Internal-only uses of generative AI are lower on the risk spectrum unless the ideas they generate are eventually presented in public.

Conduct due diligence on an AI tool’s terms and platform

In time, most of the AI platforms will be developed by companies with significant resources and standardized industry legal terms and conditions. But there are dozens of new AI tools coming out every day, many created by individuals or companies with little financial resources. They often provide no or minimal legal terms. In either case, the platforms' terms provide little protection for a user in terms of standard legal warranties, indemnities and insurance. For the moment, then, companies are using AI at their own risk. 

Determine how to handle legal liability

While AI has great benefits, its potential liabilities are unknown, so agencies will likely want to tell clients that they are responsible for the resulting risks. Agencies' position is that, for brands, the use of AI is a cost of doing business and an agency that represents many different brands cannot be the insurer behind all new technologies.

At the same time, if an agency proposes the use of AI, brands will likely argue that the agency must stand behind any vendor it brings to the table. Given that most agency-client agreements don’t directly address issues such as AI, both parties should decide how they should be restructured going forward.

Have creative AI output undergo legal review

While generative AI may be an excellent way to create new concepts or even finished products, they still need legal review.

Generally, the inspiration for content from a team of creatives can be compared to the finished product to determine whether the final product is clear or an infringement. The difficult thing about conducting a legal review on AI-generated materials is that it is not known what inspired the content created, thereby presenting inherent risk.

That being said, there are some tools that can be used to help. For example, a finished image can be run through Google’s reverse image search—to see if it is too similar to something else that’s already out on the internet. If you are creating models via generative AI, there are platforms that will allow you to compare your models against pictures of similar real people that exist on the internet.

Notably, while it was initially thought that AI-generated content might run counter to Google’s search engine guidelines, its rules don’t outright prohibit AI-generated content, but rather prohibit only “spammy automatically-generated content that has been generated programmatically without producing anything original or adding sufficient value.” 

Set a baseline for your employees, agencies, vendors and freelancers

Obviously, AI will be used in the creation of advertising materials going forward; we cannot put the genie back in the bottle. But brands and agencies must be sure to first evaluate possible risks and disclose its use in marketing materials.

So, a baseline rule for all should be that AI cannot be used without disclosure and approval. This rule applies to anyone working on behalf of a brand or agency, including employees, vendors, freelancers and other independent contractors—who might be using these technologies without disclosing them, thereby exposing companies to unknown risks.