AI Slip-Ups in Supply Chains: What You Need to Know

A big IBM report found that 13 percent of companies had a breach involving AI models or apps. In nearly all those cases, 97 percent, there were no proper AI access controls. When workers used hidden AI tools, known as shadow AI, the average breach cost jumped by 670,000 dollars. This is not just a warehouse problem.

A survey of supply chain workers found that one in six had seen AI-related data exposure. About 63 percent said a supplier or customer asked how their data was protected. For 30 percent, those questions changed which software provider, vendor, or AI tool they chose.

Most workers, 87 percent, said they trusted their AI tools to keep supplier and customer data apart. Still, 63 percent admitted some overlap existed between the two data types on their platforms. The top obstacle, picked by 22 percent, was that current AI tools were not built for this separation. Cost followed at 21 percent, and missing technical skill at 20 percent.

Nine out of ten people said they would pick an AI tool made for distribution if it offered stronger data controls. Also, 92 percent said data separation was a high or medium priority for the next year. Experts recommend testing AI with normal tasks, such as writing a customer quote without revealing secret supplier prices. They also say a person should approve AI actions before they take place.

AI tools now handle sensitive business details like customer pricing, supplier contracts, and rebate data. A leak could damage relationships with trading partners. It could also influence which technology vendors companies decide to hire.