Experts Address What ‘Huge Gap’ in AI Policy Means for Apparel Brands

President Biden’s October executive order related to artificial intelligence does not directly affect brands and retailers in the fashion and apparel space. Instead, it outlines requirements that large technology companies, those supplying the U.S. government with technology, and healthcare providers must follow.

But brands and retailers shouldn’t act as though they’ve been totally let off the hook, experts say. Regulations related to privacy and data could be ahead in 2024.

More from Sourcing Journal

Stacy Marcus, partner with New York-based law firm Reed Smith, said the responsibility of supervising AI likely will fall to agencies like the Federal Trade Commission (FTC); the executive order, she said, doesn’t offer much in the way of actionable guidance for brands.

“The executive order, while a great headline that shows the American people that the federal government is watching and concerned about these sorts of technologies… puts in these standards largely that are not applicable to brands,” Marcus told Sourcing Journal. “I think that the real, meaningful guidance will ultimately come from regulators—either agencies or legislatures—and hopefully agencies because they have a little bit more time and energy and are a little bit deeper into a specific category.”

That, she she, is partly because it would be extremely difficult for the federal government to hand out one solid set of regulations around AI due to its vastly different use cases in different industries.

Regulatory entities like the FTC have already begun evaluating AI’s impact on consumers and are working to provide information and guidance.

In October, the FTC held a roundtable to learn more about AI’s impact on creative fields; in an August blog post, it outlined some tips for companies offering consumer-facing digital products that integrated AI.

Ilana Golbin Blumenfeld, PwC’s responsible AI research lead, said the FTC and other government agencies may outline regulations related to privacy, transparency and accountability and oversight in the future. Though nobody has a crystal ball on what could be coming brands’ way, she said, the government has a track record of taking a hard look at privacy concerns.

“The government has historically had a very strong focus on consumer protection as a core tenant of a lot of their initiatives. The Blueprint [for] an AI Bill of Rights that came out several months ago, before the executive order, also reaffirmed that consumer protection, privacy, those types of attributes,” she said. “Organizations that are using AI systems really could be looking at their existing practices around privacy and the practices that could potentially impact customers in a negative way, and [those] would be areas to focus on.”

And as brands continue to implement AI into their inventory systems, customer experiences, trend forecasting and more, maintaining a strong governance structure and mindful risk mitigation practices will be paramount to success with the technology, she noted.

That includes oversight of third-party providers integrating AI solutions into a brand’s systems, especially when using proprietary data.

“The most important thing from my perspective is having a clear set of roles and responsibilities. AI governance is not something that neatly falls into the purview of one existing stakeholder within the enterprise,” Golbin Blumenfeld said. “Many organizations are trying to operate AI governance by committee, which is a great way to get started—to get voices involved from different perspectives from different facets of the enterprise—but you do need to have accountable parties and responsible parties for different portions of the risk management and data governance programs.”

Data from Aptean shows 38 percent of fashion and apparel brands are using AI in their operations, with an additional 27 percent saying they’re investigating how it can be implemented.

The data shows brands have most interest in the following use cases: data analytics, decision-making and predictive analytics; customer service; manufacturing; automation and supply chain management.

But generative AI could later help designers create styles with greater speed. Currently, Marcus noted, a lack of regulation on copyright procedure could be hindering brands’ interest in using generative AI for design.

The U.S. Copyright Office is currently collecting feedback on the connection between copyrights and AI until Dec. 6. Earlier this year, a district court upheld the office’s refusal to grant copyright to works that have been 100 percent generated by AI. The office has said that creative works created by AI alone cannot be registered under copyright, but it’s possible that AI-assisted human work could be.

The lack of clarity around what can and cannot be registered creates an issue for brands interested in using generative AI to bolster their design processes, Marcus said.

“I think that the problem for most brands is there are very few things that are completely 100 percent AI generated, and then there’s a ton of stuff in this gray area where the question remains as to how much human involvement is required in order to obtain protection and as a brand,” Marcus said. “You want to make sure that whatever you’re investing in, is protected or protectable. I think that that’s a huge gap, and one where the policy gives us no directive; the laws give us no directive.”