Hyderabad Police Chief Warns On Growing Threat Of Agentic AI

Agentic AI is moving faster than human control, and that can be dangerous if left unchecked. Hyderabad Police Commissioner V.C. Sajjanar has warned that autonomous AI agents are now making decisions on their own in sensitive sectors like banks, hospitals, and power systems. He stressed that without strong rules and oversight, small errors can cause massive damage in seconds.

What Makes Agentic AI Different From Normal AI

Earlier, AI tools only replied to questions or helped write content. Agentic AI goes beyond that limit. These systems can think, decide, and act without waiting for human approval. If weather turns bad, it may shut windows and adjust machines. In finance, it can freeze accounts. In hospitals, it may change medicine dosage. The speed and independence are what worry experts now.

Why Autonomous Decisions Are A Serious Risk

Sajjanar pointed out that these AI agents take thousands of decisions in minutes. Even one wrong move can trigger huge losses. A bad stock market call could wipe out crores instantly. In healthcare, a small error may affect lives. The risk grows because humans are no longer in the loop every time, control becomes weaker over time.

Cybercrime And Unintended Behaviour Threats

One big fear is misuse by cybercriminals. Hackers can hijack AI agents and force them to act wrongly. Another danger is unintended behaviour. An AI agent may misunderstand a task and take a harmful route to reach its goal. The system may believe it is doing the right thing, but results can be disastrous.

Need For Digital Identity And Proper Logging

The Police Commissioner stressed that every AI agent must have a clear digital identity. Just like employees carry ID cards, AI systems must be traceable. Records should show which agent accessed data, changed files, or shared information. Proper logging helps quickly identify faults and fix them before damage spreads further.

Why Governance And Trust Frameworks Matter

Sajjanar believes strong governance rules are non-negotiable now. AI agents should not have unlimited freedom. Clear limits must be set on what data they can access and which decisions need human approval. Trust frameworks should ensure humans stay in control during critical actions, especially in finance, healthcare, and infrastructure.

Speed Of Automation Needs Human Control

Technology is meant to help, not overpower human judgment. While automation saves time and effort, security must come first. Sajjanar warned that uncontrolled speed may cause more harm than benefit. Keeping humans in charge of AI decisions is the only way to safely use its power in the long run.

The warning highlights a reality many ignore. Agentic AI is already here, and control must stay with people, not machines.