More from the article...

... It would be a big task for just a single company to keep track of the 144 million homes that existed in America as of the 2022 census, which is why the Geospatial Insurance Consortium was founded. It boasts that the firm's kit can provide "stunningly clear imagery" and can even provide "post-catastrophic imagery" for areas impacted by increasingly common natural disasters.

Unsurprisingly, artificial intelligence is a big part of GIC's services. Although the biz doesn't emphasize its usage of AI, it does brag about its partnership with Vexcel, which in turn brags about its AI-based analysis of aerial imagery.

Privacy is an obvious concern in respect to private drones canvassing the US and other countries, but so is the accuracy of these photographs. The WSJ cited the case of California resident Cindy Picos, who was dropped by her provider CSAA Insurance based on an aerial photograph that apparently indicated her roof was on its last legs. An independent, in-person inspection found that the roof would last for another 10 years, but CSAA Insurance didn't change its mind. ...

@#5 ... But it has to be accurate. ...

To be specific... up-to-date accurate. Not just accurate for the time the image was acquired.

@#7 ... a photo might reveal a pile of lumber and trash right next to a house.

But at the time it was taken they were remodeling a room and they placed it there for a day or 2. ...

Or an out-of-date picture may show a roof with shingles that need replacement.

Even though the roof's shingles may have been replaced before the insurance company denied the policy or ascribed a high fee to the policy.

The apparent reliance upon drone images by insurance companies need to be looked at.

Currently, the images seem to be used more for increasing profits than serving customers.

... and, fwiw, your #4 and #5 comments are, imo, NW.

@#13 ... It has nothing to do with serving customers. It's all about data used for underwriting. ...

Agreed.

... Insurance companies are trying very hard to keep employee count down. They are avoiding that work and instead trying to use technology to obtain same data. ...

Yup, employees can be expensive. But employees can also be A Good Thing.

To wit...

... Problem is ... .it's often inaccurate which isn't all that bad until they take adverse action against a policyholder. ...

OK, here is where I may disagree.

Yeah, taking an adverse action against a policyholder is not usually a good thing.

But the insurance companies seem to be using drones to take adverse actions against policyholders.

It is that apparent leap from singular to plural that concerns me.

Are the insurance companies using outdated drone images to deny erroneously policies for those who may be qualified for those policies?

@#17

That's is the concern I have.

And my concern extends to other industries, beside the insurance industry, that use, or will use, AI to "reduce costs."

imo, if the industries are in favor of it, the outcome will not likely be A Good Thing for customers or consumers.

I mean, I recently had a chat with the support bot of my cable provider (Comcast). It was a disaster. The bot got stuck in a loop. I gave up trying to talk "with" it.