Hunton Retail Law Resource

In the past, national advertising campaigns of old required multi-day model shoots, multi-camera video productions, and multi-disciplinary graphic creation. Generative AI advertising, however, has opened a new chapter in marketing ideation and production. For retailers in particular, these tools make it faster and cheaper to create digital ads, seasonal promotions, product images, and social media content—often without involving traditional agencies or creative teams. Even old commercials and brand stories, once thought to be merely corporate history, are getting new life and are being reimagined and repurposed. While the speed-to-market and costs perspective advantages of AI generated advertising are adulated by their developers, the risks—and who bears the responsibility for those risks—are often understated. 

One such risk is that AI generated advertising raises the possibility of inadvertently using copyrighted material. Copyrighted material may be hard to detect once the final product is created, but the output can often bear a substantial similarity to the original works used to train these large language models (“LLMs”). For retailers, this creates exposure when AI-generated images, copy, or video are deployed across websites, email campaigns, or national promotions. 

Once circulated, these visual and audio works can be easily identified by copyright holders through already existing search tools, such as reverse image and audio search engines. Understanding how an advertising agencies’ or other vendors’ LLM was trained, and building proper limitations, indemnities, and clearance into master services agreements with those third parties, is essential. Taking these necessary precautions are important since in all likelihood, the retailer, not the third party, will be subject to an enforcement action. 

But it’s not just a matter of avoiding liability. AI tools format their responses based on a combination of their existing LLMs and the prompts provided to them. When prompting a generative AI tool to create what is thought to be “unique,” it’s important to protect the skill, costs and consumer goodwill invested in that final product. Copyright law requires human authorship to qualify for protection, and generated artistic works, without meaningful human involvement, generally are not subject to protection. 

Just like an inventor is advised to save her lab notebook under patent law, maintaining logs of sketches, prompts, and edits is important for proof of authorship and exclusivity—a practical step for retailers seeking to protect brand assets created with AI assistance—and the level of human involvement will be a significant consideration. Recognizing that many generative-AI tools consider user prompts not to be confidential is important when considering how to maintain exclusivity. 

Trademark issues also warrant attention. While copyright issues tend to dominate the conversation about AI training data, there are additional risks under trademark law. When an AI system generates output, for example, it often references or imitates protected brand names, logos, or slogans. This can be especially problematic in retail advertising, where brand identity and source‑identifying elements are critical. Therefore, AI advertising creators must check for potential trademark infringement in their content by clearing key names, logos, and design elements before materials go live in consumer-facing campaigns. Doing so allows creators to preemptively edit their AI prompts and have peace of mind. 

The right of publicity and privacy considerations are equally important. When generating advertisements that use images of people or their personal data, creators must secure protections against unauthorized use. Retailers using AI to generate “models,” influencers, or lifestyle imagery should be particularly cautious where images resemble real individuals. To that end, it is important to understand the AI platform’s applicable disclosures and identify whether right of publicity and privacy protections are already in place. If an ad involves or resembles an identified person, retailers should consider whether it is appropriate to obtain a release or permission. Failure to do so can lead to legal action and reputational harm. 

More broadly, the speed and ease of AI‑generated advertising should not replace traditional legal review and brand‑protection guardrails. Creating a company-wide procedure that identifies the requisite terms, tools, and permissions that are required before using generative AI for advertising is crucial for long-term success and can help retail legal teams balance innovation with risk management. Preventative processes will help educate employees on the issues, promote consistency across marketing teams, and reduce the likelihood that AI-generated content creates unintended legal exposure.