Want to learn more about AI in person? Check out ODSC AI West 2026, coming to Burlingame this October 27th-29th: https://hubs.li/Q04cYsmk0 #DataScience #AI #ArtificialIntelligence #ODSCAI ------------------------------------------------------------------------------------------------------------- Visit our website and choose the nearest ODSC event to attend and experience all our training and workshops: https://odsc.ai To watch more videos like this, visit https://aiplus.training Sign up for the newsletter to stay up to date with the latest trends in data science: https://opendatascience.com/newsletter/ Follow us online! • Facebook: https://www.facebook.com/OPENDATASCI • Instagram: https://www.instagram.com/odsc/ • Blog: https://opendatascience.com/ • LinkedIn: https://www.linkedin.com/company/open-data-science/ • X (twitter): https://x.com/_odsc

#CiscoSecurityPartnerYour AI agent has a login. Nobody's checking its badge. Everyone's racing to ship agents. Almost nobody can answer the one question that matters: what is this thing actually allowed to touch, and who stops it when it goes rogue? That blind spot is where the next breach lives. The field data says it out loud: 85% of enterprises are experimenting with agents. Only 5% made it to production. And 59% of security leaders say the agents themselves are the wall they can't get past. Read that again. The blocker isn't the model. It's control. Here's what most rollouts are quietly ignoring: An agent is an identity. It holds tokens. It calls APIs. It chains actions across your stack faster than any human, with more access than any human would ever get. We spent ten years enforcing least-privilege for people, then handed agents long-lived, over-scoped tokens and called it innovation. And the failure modes are brutal: → Prompt injection isn't a content bug. It's privilege escalation. Untrusted text becomes a command, and it runs with your agent's access. → One compromised agent + tool-calling = lateral movement across everything it can reach. → Auth once at login? Useless when the agent drifts mid-task. You approved the intent, then stopped watching. You can't fix this with a dashboard bolted on after launch. You fix it where the traffic actually flows, the network. That's why Cisco Security's approach holds up. Cisco SASE turns the network into a live control plane for the agent era: identity-aware policy for non-human actors, continuous inspection instead of trust-at-login, real control over what agents can reach, converged with networking so it adds zero drag. Dual defense, plainly: keep hostile agents out. Stop your own agents from becoming the incident. Least-privilege was optional for humans. For agents, it's survival. The winners of this era won't have the most agents. They'll be the only ones who can prove, in real time, exactly what every agent is doing. Can you? #CiscoSecurityPartner #CiscoSecureAccess #AgenticAI #ZeroTrust #SASE #NHI #TheRavitShow