Retailers are facing AI risks, industry lawyer warns

This article was originally published by Christopher Kelly for Ragtrader (2 July 2024).

Retailers jumping into the use of generative artificial intelligence (AI) for selling products or connecting with consumers need to be wary of legal and reputational risks. 

This is according to Bartier Perry commercial partner Rebecca Hegarty, stating that many retailers weren’t prepared or fully aware that the use of generative AI had the potential to backfire badly on a business. 

“Reports indicate the generative AI market will be worth an estimated $1.3 trillion by 2032,” Hegarty said. “So, it’s not surprising many retailers are adopting or trialling the technology, but they should be aware of the risks. 

“Simply claiming that it was the AI’s fault when something goes wrong is neither a sound legal or reputational defence.” 

Retailers are beginning to use AI models to allow for more realistic online try-ons of clothing, all the way through to personally curating customer experiences. 

Ragtrader also knows of fashion retailers using AI to drive marketing campaigns through the use of customer data.

Last year, a New Zealand supermarket’s AI-driven meal planner for leftovers began suggesting recipes such as Oreo cookie vegetable stir-fry, with users then highlighting other AI suggestions included ant-poison and glue sandwiches and bleach-infused rice surprise. 

“These so-called AI hallucinations may or may not be covered by legal disclaimers, but either way they risk brand damage,” Hegarty said. 

“From a purely legal perspective, putting customer or other data into AI tools or models raises significant data protection and privacy issues. There are also risks in retailers unwittingly inputting or assigning valuable commercial or intellectual property to a generative AI tool.” 

Hegarty said retailers need to ensure an appropriate AI governance framework or AI risk mitigation action plan is in place to assess the unique risks of the technology. 

Her key points to consider are what data the AI tool is being trained on, whether employers understand the risks involved, which cybersecurity safeguards are in place and whether employees can check outputs for hallucinations, errors, defamatory or plagiarised content.

“AI is not a set and forget technology,” Hegarty said. “Retailers will need to be sure they keep across it and check-in as to how it is interacting with their customers in the same way they do with their retail team members.”