{"id":541214,"date":"2026-03-23T22:56:11","date_gmt":"2026-03-23T22:56:11","guid":{"rendered":"https:\/\/www.newsbeep.com\/us\/541214\/"},"modified":"2026-03-23T22:56:11","modified_gmt":"2026-03-23T22:56:11","slug":"cisco-launches-security-services-to-protect-the-world-from-ai-agents-and-each-other","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/us\/541214\/","title":{"rendered":"Cisco launches security services to protect the world from AI agents and each other"},"content":{"rendered":"\n

OpenClaw, the popular AI agent platform, is a massive hit among the techy set. Developed by Peter Steinberger, who has since joined OpenAI (OPAI.PVT<\/a>), OpenClaw allows users to set up their own AI agent that can run on their computers, performing tasks ranging from checking email and replying to messages to managing system files.<\/p>\n

And while those capabilities are impressive, they\u2019re also a potential security nightmare. That\u2019s because you\u2019re giving an AI model the ability to control your computer. And while that\u2019s great when it works, it can lead to major issues, whether that\u2019s permanently deleting important emails or entire programs from your system.<\/p>\n

And that becomes an even bigger problem for companies testing OpenClaw on their enterprise devices.<\/p>\n

To that end, a number of organizations are stepping up with new security offerings designed to corral OpenClaw, helping to keep its AI agents from becoming more trouble than they\u2019re worth.<\/p>\n

\"The OpenClaw logo on the software’s website, March 12, 2026. (REUTERS\/Florence Lo\/Illustration) \u00b7 REUTERS \/ REUTERS <\/p>\n

Nvidia (NVDA<\/a>) debuted its own such platform during its GTC event in San Jose, Calif., last week called NemoClaw,<\/a> and on Monday Cisco (CSCO<\/a>) showed off its entry into the space.<\/p>\n

\u201cOpenClaw is that ChatGPT moment for agents, and it’s actually a catalyst for making it easier for tech-savvy users to deploy agents at work, at home,\u201d explained Cisco president and chief product officer Jeetu Patel.<\/p>\n

\u201cThe downside that is a challenge \u2026 is this risk around trust and security for these agents. Because these agents \u2026 they’re kind of like teenagers. They’re supremely confident, they’re inexperienced, they don’t fully appreciate the consequences, and they have no idea where the rules exist,\u201d Patel added.<\/p>\n

To address those concerns, Cisco says it\u2019s taking a multipronged approach that seeks to protect the world from rogue AI agents, protect AI agents from the world, and use AI to detect potential threats from malicious actors and hackers.<\/p>\n

Defending the world from AI agents sounds hyperbolic, but according to Patel, it\u2019s not. Agents, like any piece of software, can be exploited to attack infrastructure or steal sensitive data.<\/p>\n

\u201cThe stakes of these agents are much higher than [with] chatbots. With chatbots, the worst-case scenario is that you get the wrong answer. With agents, the worst-case scenario is you actually encounter the wrong action from the agent, and some of these actions could be pretty catastrophic, like emptying your bank account or deleting your email, which can’t be undone,\u201d Patel said.<\/p>\n

Cisco, like other companies, says the best bet is to treat AI agents like you would human workers, giving them verifiable identities that you can then manage and providing the digital helpers with the appropriate access to the right systems and services.<\/p>\n

Protecting AI agents from the world, meanwhile, is meant to prevent hackers from jailbreaking the AI and forcing it to perform tasks that it otherwise wouldn\u2019t.<\/p>\n

To do that, Cisco says it\u2019s launching a service that will allow customers to test their AI agents against simulated attacks to see where they fare well and where they come up short.<\/p>\n

Finally, the company says it\u2019s leaning on agents to respond to other AI threats, ensuring that customers can react to instances with the same kind of speed as attackers.<\/p>\n

OpenClaw is just the latest permutation of the AI explosion and is likely to drive further changes in the near future. After all, it only debuted in November and is already driving business decisions at major corporations.<\/p>\n

That, of course, will necessitate even more types of cybersecurity technologies and techniques to help defenders catch up to attackers.<\/p>\n

\"Sign<\/a> Sign up for Yahoo Finance’s Week in Tech newsletter. \u00b7 Yahoo Finance <\/p>\n

Email Daniel Howley at dhowley@yahoofinance.com. Follow him on X at @DanielHowley<\/a>.<\/p>\n

Click here for the latest technology news that will impact the stock market<\/a><\/p>\n

Read the latest financial and business news from Yahoo Finance<\/a><\/p>\n