Businesses rush to lock down AI as autonomous agents reshape workplaces
Businesses are scrambling to secure AI systems as new risks emerge in the workplace. With autonomous agents becoming more common, companies face growing pressure to prevent security gaps. Recent moves by tech giants and regulators highlight the urgency of the challenge.
At the RSA Conference 2026, the focus has shifted to protecting the 'agentic workplace'—where AI tools operate with increasing independence. Meanwhile, studies show nearly all cybersecurity leaders have delayed AI projects over safety concerns, exposing a widespread 'readiness gap' in organisations.
The push for tighter controls comes as one in eight firms now reports security incidents tied to AI agents. While no confirmed breaches from autonomous agents like OpenAI's Operator have been documented, broader fears persist. The EU AI Act has raised liability risks for businesses, forcing many to rethink compliance under stricter rules.
In response, Cisco unveiled *DefenseClaw*, an open-source framework that scans and sandbox-tests every AI agent capability before it acts. The company is also embedding agent identity management into its *Duo* platform, letting firms register AI tools as distinct identities under human oversight. CrowdStrike, too, is expanding its *Falcon* platform to detect unauthorised AI applications and shield the systems running large language models. Regulators are stepping in as well. The U.S. government released a national AI framework to standardise federal security measures and avoid conflicting state laws. California, meanwhile, is refining legislation to exclude autonomous agents from the legal definition of a 'person' in public records. On the software front, Microsoft plans to roll out *Agent 365* on 1 May 2026. The update will embed autonomous agent features into Office applications, complete with new security dashboards for monitoring.
The race to secure AI-driven workplaces is accelerating, with tools like DefenseClaw and Falcon aiming to close vulnerabilities. Regulatory shifts and upcoming releases such as Agent 365 will further shape how businesses manage risk. For now, the focus remains on bridging the gap between AI adoption and robust safeguards.
Read also:
- India's Agriculture Minister Reviews Sector Progress Amid Heavy Rains, Crop Areas Up
- Sleep Maxxing Trends and Tips: New Zealanders Seek Better Rest
- Over 1.7M in Baden-Württemberg at Poverty Risk, Emmendingen's Housing Crisis Urgent
- Cyprus, Kuwait Strengthen Strategic Partnership with Upcoming Ministerial Meeting
