Okta CEO Todd McKinnon has raised serious concerns about the security risks being unleashed by artificial intelligence as it leaps out of experimental labs and into real-world business applications.
In a recent interview, McKinnon told CNBC’s Jim Cramer that robust identity-based security systems will be necessary to manage the growing risk associated with live environments and artificial intelligence.
AI deployment raises new security challenges
As AI agents move from test environments to actual roles, they will have real-time access to highly sensitive systems that include confidential customer information, internal files, and business data, said McKinnon. That is similar, he said, to bringing AI into production, which is akin to bringing in new coworkers you’re going to monitor closely.
If not carefully protected, these systems could automatically make decisions that wind up leaking data, allowing unauthorized access to that information, or using it in unintentional ways. Companies need advanced monitoring systems that monitor and control the deployment of AI agents so that AI agents only possess access to the system based on a trusted and authorized identity, said McKinnon.
Okta’s approach to AI security and growth strategy
Okta logged stronger-than-expected earnings and revenue for the recent quarter but stuck to a conservative outlook for future growth. That sent the stock tumbling over 11% in after-hours trading. That’s not the kind of growth you can forecast confidently in an unpredictable global economic environment, McKinnon said.
He said Okta’s focus on stability will supersede aggressive expansion. It plans for the long term with slow, steady growth to defend its position in the market. He also said that chucking up identity systems used by both human users and AI agents could cause operational errors.
Industry warnings as Okta prepares for an AI-driven future
McKinnon’s concerns have been echoed by industry experts, warning that the deployment of advanced AI, like Autofac, takes precedence over speed. Last month, Max Tegmark, a leading voice in AI safety, suggested that the development of AI systems that are stable enough could have a high probability of posing existential risk, similar to how risk calculations were made before the first nuclear test. However, balancing AI adoption with sober identity-based security protections is crucial, according to McKinnon.
Okta is ready to lead in this space by providing scalable solutions for secure human user and AI agent access management, he said. McKinnon pointed out that as AI technology develops, organizations need to ensure careful planning, patience, and discipline to remain resilient. Okta is optimistic about the future, but they’re focused on building systems that are safe and secure in an AI world that’s only going to become more AI and more technology.
The post Okta CEO Issues Warning on AI Security Risks first appeared on Coinfea.