Mitsui Sumitomo Insurance Group

Today's guest is Prashant Hinge.

He is the Chief Information and Transformation Officer at MSIG USA.

And it's another interview that I recorded in a cafe in New York last week.

Joining me to have a quick chat before we get into the interview is Sam.

Yes.

Tell us a little bit about...

MSIG and the role that you do within the business?

Sure.

This is very different from a human-led attack, but it still has the potential to cause a lot of monetary damage.

Insurers, including MSIG, QBE, and Beazley, are reviewing how their policy language applies to these scenarios and who bears responsibility when an agent's autonomous actions cause damage.

For now, most of them are clarifying the parameters of their coverage rather than outright excluding AI events.

QBE's global head of cyber described AI as a, quote, "risk amplifier, not a fundamentally new cyber risk." End quote.

We're also going to see agentic liability insurance explode.

MSIG and Beasley are already rewriting policies around agents escaping test environments.

That Aurora breach, where the agent's own reasoning trace is the smoking gun, will become the reference case.

Whoever writes the actuarial model for agent authority sets the de facto governance standard way ahead of any actual regulator.

Second, re-evaluate your enterprise cyber insurance and liability coverage without delay.

With major carriers like MSIG and Beazley already actively restructuring policy terms around AI agents escaping designated test sandboxes, extending autonomous agent authority into physical hardware without explicit updated underwriting creates a dangerous liability exposure.

Third, adopt the rigorous stress testing protocol recommended for autonomous coding agents and implement it across any hardware integrated system.

Before granting an agent live operational access to physical machinery, run adversarial, this is just a test bypass prompts against its safety boundaries and patch the governance gaps before bad actors exploit them.