3 AI Experts Urge Clearer Governance and Human Accountability for Agentic Systems
Updated
Updated · ZDNet · Jun 25
3 AI Experts Urge Clearer Governance and Human Accountability for Agentic Systems
1 articles · Updated · ZDNet · Jun 25
Summary
Vint Cerf, David Bray and Cheryl Einhorn said companies need clearer rules for how humans and AI agents work together, arguing trust in AI depends on precise instructions, human judgment and defined accountability.
Cerf warned that AI agents using natural language could misread one another and act "at the speed of light," while Bray likened today’s AI environment to 1910 streets before stoplights and right-of-way rules.
On trust, the panel said AI should be treated less like an oracle than a colleague whose output must be checked; Bray urged triangulating results, and Einhorn called AI adoption a broader cultural shift in problem-solving.
On accountability, Einhorn said humans still bear the consequences when AI fails, Bray asked whose organizational "flag" an agent flies, and Cerf argued businesses need formal recourse when systems cause harm.
Bray said more than 40% of the world’s information could be synthetically produced by 2030, raising governance and verification pressure for CEOs, boards and policymakers.
The EU's AI accountability law starts in August. Are tech companies and their users truly prepared for the consequences?
When we must constantly second-guess our AI 'sidecar,' are we saving time or creating new cognitive burdens?
With 90% of online content soon to be AI-generated, is authentic human thought becoming an endangered species?
Agentic AI’s Autonomous Surge: 72% Enterprise Use, 60% Governance Gap, and the Critical Trust Imperative
Overview
Agentic AI marks a major shift in artificial intelligence, moving beyond systems that simply process data or generate content. These advanced systems are designed to operate with a high degree of independence, setting their own goals, making plans, and executing tasks in dynamic environments without constant human oversight. Agentic AI can perceive its surroundings, reason about situations, learn from experience, and adapt its behavior to achieve objectives. This ability to act autonomously and pursue complex, multi-step goals raises important questions about control, accountability, and safety, making robust governance and oversight more urgent than ever.