Lupo and Marsh Quoted on Digital Risks of AI

“Fashion companies are particularly exposed because they’re not banks – they’re not staffed or regulated for this level of digital risk.” Tony said. “They often don’t have the budget or the technical infrastructure to protect themselves the way financial institutions do.”

Tony noted that the risks begin when choosing a reputable vendor and avoid contracting with unreliable AI systems. “Some vendors are offering deep discounts in the first year, but once you’re in, you’re stuck – and the costs skyrocket later.”

But more traditional risks persist.

“AI will take traditional attacks – like ransomware, domain hijacking, and supply chain disruptions – and elevate them to a much more sophisticated level,” Tony said. “I’m very nervous for companies that are still in what I call ‘stage one,’ the early awareness stage of AI experimentation.”

He added, “As amazing as these AI systems are, they’re also going to make it easier for bad actors to perpetrate fraud and gain access to systems in ways they never could before. The only real chance companies have to protect themselves is to use AI to combat AI.”

Tony said that brands must plan a defense as well as an offense. “Every brand needs a SWAT-style plan for what happens in a breach – not just IT, but finance, HR and legal. You can’t only focus on AI’s creative or cost-saving potential; the defensive side deserves equal attention.”

Human oversight, bolstered by outside counsel, may be a brand’s best defense.

Michelle said that disclosure and generous returns policies will be brands’ best short-term friends while consumers adjust to AI-augmented fashion retail experiences. As regulations evolve, brands must be ready to demonstrate not only that their systems are effective, but that they are fair, explainable, and safe.

Tony added, “I do think we’ll see fashion companies appoint chief AI offices. Part of their responsibility will be brand and data protection, but most CEOs will expect them to focus primarily on growth and efficiency – and that imbalance creates risk.”

“Pandora’s box is open – you can’t avoid using AI,” he said. “But you can control how you use it. Keep financial and customer data in a sandbox, separate from generative tools, and make sure your systems are insured and auditable.”

Read the full article here. (Subscription required)