Stephen Nellis
(Reuters) – Trend Micro said on Sunday it was working with Nvidia to develop new cybersecurity tools that use artificial intelligence and are designed to protect the data centers where AI runs.
The tool, which Trend Micro plans to unveil at its Computex conference in Taiwan starting Sunday, can run on Nvidia chips and is designed to detect intruders and ensure that only those authorized to use the data can see it.
Many companies are training AI systems to assist employees with tasks like answering HR questions or assisting customer service agents, but to do so they often collect data from across the enterprise in one place, making the systems an attractive target for hackers.
“They go inside a company and find huge honeypots of information,” Trend Micro Chief Operating Officer Kevin Simser told Reuters.
In addition to detecting intruders, another area of focus for Trend Micro is ensuring that hackers cannot snoop on the data being fed into AI systems. With most modern chatbots, users interact with the bot by asking it questions, and the questions themselves often contain sensitive information, such as undisclosed trade secrets or personal customer data.
“They're often narrowing the scope[of the chatbot's response]by providing very specific information,” Simser said of using AI prompts. “That's what we're looking for, to make sure we review it first and don't go any further than the people or applications that are authorized to see it.”
(Reporting by Stephen Nellis in San Francisco; Editing by Cynthia Osterman)
