AI risks make some insurers wary of corporate liability

If AI does end up killing us all, or just compounds our daily havoc, the insurance industry would rather not pay the bill. "AI-related harms are already emerging, including incorrect or misleading outputs, deepfakes, privacy violations, intellectual property disputes, fraud, product defects, and discriminatory decisions," …

[Anonymous Coward]

The apocalypse has been prevented ..

.. by bookkeepers?

Now that's something I didn't see coming.

[DS999]

They were already absolved from insuring the really bad consequences

Because they have policy limits in the small billions, not the hundreds of billions or larger that an AI going apeshit and hacking into stuff that directly results in the death of a lot of people would cost.

But it is highly interesting that they shying away from insuring companies who are merely USING AI. That's new to me, but is really going to hurt the revenue side of the ongoing very very very very costly AI buildout binge. Most boards will think long and hard about wide deployment of AI - even if it in theory saves a lot of money by letting them fire a bunch of people - if they are concerned that if the AI does something wrong their insurance is not going to pay out and it will instead essentially come out of the stock value.

[Cubbie Roo]

The mask slips

https://www.404media.co/doom-loop-openai-and-microsoft-admits-llms-are-destroying-the-web-and-built-on-theft/

[Groo The Wanderer - A Canuck]

Risk Analysis and Costing is going to cost any business or individual who even touches an Artificial Ignorance system of any kind is going to find their premiums going up according to whatever amortization tables fall out the end of a few years of cost collection and risk analysis optimization as to which factors are most relevant to impacting risk, and by a quantitatively predictable amount, if lucky. If not, the worst case plus 10% will do...

[jake]

Real World consequences .

I have called a hakf-dozen malpractice insurers here in the SF Bay Area, asking if my GP was still covered now that he's started using AI in his day-to-day practice, including diagnoses.

The universal answer was "absolutely not", followed by "he should really read the policy, it's in there".

He's pissed at me now, and I need to find a new GP ... No great loss.

[Claptrap314]

Re: Real World consequences .

That's actually great news. We need to press on this. They're using LLMs to read x-rays for crying out loud! I can believe that NNs could eventually get to that level, but LLMs?

[Claptrap314]

Isurance company saviors, round II...

You may remember that a decade ago, some of us were holding out hope that the insurance companies would force an industry turnaround, as the quality of software was becoming so bad. That appeared to fall through when they continued to cover paying attackers for data recovery.

So I'm not holding out much hope for a better result this time. Would be nice, however.

[amanfromMars 1]

Quite a Horny Dilemma indeed. Damned if You Do, Damned if You Don’t.

Companies want to protect themselves from the financial liability of deploying unreliable, mendacious, amoral AI agents, but insurance companies have become wary about taking on that risk.

:-) Oh dear. What a shame. Whatever next. And However, if the incalculable risk from the use of AIgents ends up being uninsurable and problematical, the insurance industry will quite naturally and logically become a prime target for AIgent remediation ..... specialised particular and peculiar attention .... and that is sure to be extremely uncomfortable for any and all insurance industry parties and partners worthy of such unusual and novel undivided attention.

[jake]

Re: Quite a Horny Dilemma indeed. Damned if You Do, Damned if You Don’t.

Eh? Are you seriously suggesting that AI is going to put paid to the insurance industry?

I hate to say it, my friend, but I rather suspect yer butter's done slipped off yer biscuit.

_{Perhaps have a beer and re-think what you are saying? I'm buying.}

[amanfromMars 1]

Re: Quite a Horny Dilemma indeed. Damned if You Do, Damned if You Don’t.

I agree, jake. The opportunity to continue to make easy money off the AI Money Generation Machine is just far too tempting for the insurance industry to resist and deny insurance cover for ...... no matter how expensive their algorithms tell them to make it. :-)

But ..... to underestimate what AI, even in its current relative infancy form can do, leaves one catastrophically vulnerable to that which it can surprisingly easily and frighteningly quickly do ....... with the buggering up of the insurance industry probably one of the least of everyone's existential worries.

Cheers for the beer. The next one's on me.

[Tron]

It's clearly serious.

On a good day, insurers can dodge paying out for almost anything. If they want clear blue water between the unlimited amount of expensive crap that a single agentic spree can cause a company, then that is the largest red flag you are going to see outside Beijing.

RAND are actually being quite cautious, so as not to enrage the orange goblin, who needs AI to keep faking the economy with circular investment.

If I was an insurer, I wouldn't insure any company that didn't keep its data/intranet on prem, using a second network for internet interaction. If your stuff is connected to the net, and it is worth hacking, it is only a question of time before hackers armed with AI and newly publicised zero days add you to their list of victims. Connecting any network with anything of value on it to the internet, thanks to AI, may no longer be an option. AI has become the hacker's tool par excellence and they are seeing the fabled RoI that nobody else is.