https://www.businessinsider.com/homeowners-insurance-nightma...
Cahn is the founder and executive director of the Surveillance Technology Oversight Project, or STOP, a New York-based civil-rights and privacy group, so he certainly has a dog in the ring here, but he also has a horror story to back it up.
The source article on Business Insider contains much more important details:
> Travelers admitted that it screwed up. It never conceded that its AI was wrong to tag me. But it revealed the reason I couldn't find my cancellation notice: The company never sent it.
> Travelers may have invested huge sums in neural networks and drones, but it apparently never updated its billing software to reliably handle the basics. Without a nonrenewal notice, it couldn't legally cancel coverage. Bad cutting edge tech screwed me over; bad basic software bailed me out.
So basically, this comes down to a dispute over how much moss is too much moss to make a roof structurally unsafe. But it sounds like the process goes straight from "AI detects a problem" to "policy gets cancelled," without human review in the middle. Perhaps a less error-prone way of handling it is for the AI's recommendations to trigger a human to go out to the home and investigate?
But, how can this be outsourced overseas ?
The entire story: His insurance broker said his policy was cancelled because of AI. His policy was never actually cancelled.
Everything else is complete speculation.
The irony is that the most likely culprit in all of this was simple human error with his insurance broker.
Image recognition is not really "AI-driven", and the low numbers make that replaceable with humans. It's the cost and legality of drone roof photos that make this possible.
The risk represented in the photos was relatively small, but it's a risk easily and legally measured. Then the higher cost of fix + verify is shifted to the homeowner.
The real beneficiary is roofing companies, raising the question of illegal tying. Insurance is required by mortgages, so homeowners have no choice but abate with roofing services, which creates an opportunity for the insurance company and roofers to share value extracted in various ways. Which ways are legal is an open question. The value extracted is bounded by the cost of switching, which involves another company assessing your property in some way; tight home insurance markets thus increase the value extractable.
Insurance mandates and reliance require regulation, as does using private insurance for large social risks like wildfire and earthquakes, but that all makes insurance less competitive by reducing viability of new entrants.
Nothing in the chain of reasoning - from drone pictures to investor decisions - is improper, but boy the resulting homeowner squeeze is painful.
If the insurance company convinced you to go up and "remediate" the moss problem it sounds like the system worked and they spared you (and them) from an expensive roof replacement.
This opportunity is naturally enlarged in jurisdictions that drive down the population of business competitors with aggressive misregulation -- California.
I would expect the insurance company would say something like, "We noticed this problem with your roof. Please correct it or we'll cancel your policy." And then the homeowner is free to choose any roofing company they want to remediate.
Heck, the insurance company could even say, "Here are some reputable roofing companies in your area, but you're still free to select any company you want," and I doubt that would constitute tying.
And anyway, even if you could make the argument that roofing companies in the aggregate will benefit from these sorts of policies, it's not as if keeping one's roof in good condition is forcing consumers to make unreasonable purchases.
How is the insurance company going to get a share of the "value extracted"?
Can an auto insurer do the same and cancel my coverage because they see me doing burnouts in my driveway?
Seems like a giant privacy violation but I'm no legal expert.
It's a self-defeating cycle. As the models improve, they'll start excluding anyone who might ever need to file a claim. The more they optimize, the less relevant they become. In the end, insurance might evolve into something entirely different or become obsolete altogether. It's the ultimate paradox of trying to reduce risk to zero: You end up with no business at all.
And these applications, crucially, are not pie-in-the-sky, someday-this-will-work type situations as is comprehension of knowledge by an LLM, or a video generator that can remember what a character looks like, no. This is exactly what ML is already used for: aggregate analysis of massive amounts of statistical data. This is it's bread and butter.
That's worth $80 billion. Not shitty melty royalty-free images.
Your car insurance isn’t high value enough but yes they already do this otherwise.
Progressive is one company that insures 27 million drivers. If they deployed this tech to all of them and it earned them a paltry $2 more annually per driver, that's $54 million in extra revenue. And any company that dares not do it once it's widely adopted will have shareholders screaming at them to implement it.
No matter the source of the data, all data available is being used for any and all purposes regardless of morals or ethics. As such in business profits are the goal however in war it is to win.
As a multiple fintech founder from the 1990s I have experience with the earliest payment breaches, some known, some not. The banks were buying this data from crims to offset losses by proactively changing the impacted account numbers before they could be used. Few complained when the bank changed their account number unless of course one had a unique sequential card number. ;)
There is an ever increasing amount of public data being produced daily from countless sensors and devices now covering our entire world and to some degree space. Given this evolution one is to believe they can still hide? Just because data comes from a certain sensor or source does not limit the data use in any specific way and if you believe otherwise I have ocean front property in Kansas to deed you. Now consider the number of known breaches today since the internet went public and multiply that value by some large number N to postulate a total breaches number ever. Point being is there is a lot of data available if one knows where to look and HOW to apply that information for "profit" and/or to "win".
Just as banks were paying crims for data to proactively offset losses this mindset works in other business' today as well however it is certainly not being received openly given the impact to consumers during these increasingly challenging economic times which only greatly compounds that impact. Do recall that one can just change their financial account information but one cannot change their health records so why does that matter? Let us propose the hypothetical that health insurance companies are now buying up health records that have been leaked. We all know why they are doing this because these health insurance companies are looking to aggressively target the healthy to sign up with them correct?
Interesting times ahead as "change" continues.
Stay Healthy!
For the most part though, insurers don't actually have too much incentive to create false positives. Increasing your rates for life insurance or cancelling your policy because you ate a burger when they already have your cholesterol information is really dumb and bad for business.
This nation is built to support corporations, not personal freedom. This is the result of that architecture.
A commercial pilot's license requires 1500 hours of flight time under one's belt. The company paid them a small sum to strap a camera to their planes and fly certain routes. New pilots make some extra cash and the company gets cheap shots.
I wonder if there exists any kind of service for similarly leveraging technology to help consumers find claims opportunities. E.g. run a drone after a hail storm to look for damages that consumers could file claims with. Ideally the carrier would also be doing this, but most aren't going to volunteer money away.
It seems like whoever pays for an inspection (such as using drone surveillance), it should be shared with the other party?
Insurance companies thrive in the margins of customers being overly-conservative, and as the information asymmetry drops, their margins go down.
The drones are feeding data. It's drone-driven AI analysis.
They could cut the AI out of the loop and have humans making bad judgment calls and the result would be the same.
I would further have issues with any drone flying over my property uninvited... even if I don't strictly have issues with companies automating such inspections conditional on them being agreed to in advance.
This is clickbait.
And, thanks to our darn near unrestricted version of free speech and a Supreme Court ruling that tore down most restrictions on political lobbying by ruling that spending money also counts as a form of speech, corporations spend an obscene amount of money on getting buddy-buddy with legislators. So the legislatures also tend to be incredibly slow to act on these kinds of things because they don't want to bite the hand that feeds them, either.