@@DeepStateDestroyer yeah, no. That attitude is how you get skynet in the first place. All ai is bad AT THINGS OUTSIDE ITS SCOPE. you need to use the RIGHT ai for the job or you get shit like this.
I’m or afraid of AI, I’m just afraid of people treating like it’s authoritative and unquestionable. It’s one thing to use a specialist AI to review a big batch of imaging files and flag the ones that a radiologist should probably look at. It’s something else entirely to say “The AI decided this and we don’t know why but we have to abide by its decision without any human review!”
“AI said no, we gotta wait and see if it changes its mind” is the most ridiculous dystopian nightmare ever. But I’m pretty sure this is also illegal. They have to tell you the reason your claim was denied, and “computer was cranky today” isn’t a reason. If they refuse to resolve this, take them to court.
They will program their AI to give justification eventually I'm sure but I can't believe they are already using the AI without that behavior. Seems completely inappropriate and negligent for medical decisions. It's a huge lawsuit waiting to happen once a denial causes loss of life.
As an ex- medical biller, I can tell you that it is 100% law that they must tell you the reason for your insurance claim denial. They cannot blame not having a reasoning on AI or anything else. **Clarification** You have a legal right to take your denial to arbitration where you are allowed to question the person responsible for the denial. Many times the docs used by carriers cannot practice medicine due to a variety of issues or they are ‘out of scope’ due to their specialty. Example: an Obgyn’s denying a brain surgery. Using AI would strip the insured of that ability (at least as far as I can see from this vid)
@@brianpinkey676 very true. What a reg rep is SUPPOSED to do is inform you of your right to request ALL info pertaining to the evaluation/review of the claim/pre-auth and denial, and their particular company’s procedure for that. It’s one of the best ways to get your denial overturned actually, because you will often find that your claim was reviewed by a doctor that was ‘out of scope of practice’. For example: an Obgyn may be behind the denial for brain surgery. You can appeal on that kind of basis alone. You will be denied again, BUT this will typically be all you need to request the claim/pre-auth go to 3rd party arbitration. This is where you win because even the arbiters that are hired by the carrier cannot defend a verified diagnosis by a specialist over an out of scope doc. Also, 99% of all docs that work for carriers do it because they have so many strikes on their professional record they cannot find work practicing medicine. All of that info is public record. So when you have 1-2 respected neurologists in your favor, it’s kinda easy to make the argument that their say as to the need of a procedure/med/etc should be held in higher regard than an OBGYN with several malpractice settlements and a suspension for a positive drug test on their record. In fact i’m shocked a carrier would use AI, as the laws have not kept up with tech. You have rights in arbitration that are impossible to exercise with a computer making these decisions
Perfect way to ensure profitability: Make the AI deny more claims over time and then just say "well the computer said so". This is a MASSIVE lawsuit waiting to happen
@@EnragedSephiroth that's what ai generally is, it's not intelligent. It doesn't understand nuance and individual circumstances, it uses a set of things that have already happened and if it doesn't fit that set it will probably deny.
What's actually happening is there is a quota, when it's not filled the ai is more forgiving. Another thing is ppl are lying to get past the AI. AI is just a calculator 1 + 1 will always equal 2
Well, it's how all LLMs (text generation AIs) work. It's the same reason why you can ask chatgpt the same thing twice and it will reply you totally different stuff
@@MAJ4KIt depends on the AI being used. Some models have some random chance thrown in, meaning you can get different outputs when giving the exact same inputs. This is fine for something like image or creative text generation, not so good for cases like these when you need a deterministic answer! Judging by this video I wouldn't be surprised if the model used here was non-deterministic.
I guarantee the prompt given to the AI includes the phrase, "Find a reason to deny this claim." OP is probably in good health and the AI said, "No reason for annual physical"
Why would providing a reason make it any more fair when they can just make up a real reason? Where is it their job to be fair and honest with their costumers? There are illegal things that they can’t do with your coverage, but they aren’t hand tied to provide good insurance either. It’s still worth it for most propel, even if exploitative and risky for healthcare in general
@@yucol5661 because when you state a specification it can be challenged and is set in stone. If the issue is nebulous it adds a lair of obfuscation so you first have to go through the hoops to get it defined first and THEN you can argue about the reason. It legalese for were gonna tey an bog you down in red tape to hope you just drop it.
"You will be rewarded 15$ for every claim you deny! Try to deny as many as you can to optimize your earnings. It's early October, a very productive period."
This is so great for these companies. In addition to not having to pay anyone to look at these claims, they have found the holy grail of deferred responsibility. “Sorry but for reasons unknown our black box computer said you don’t deserve coverage” you can get away with anything if no one is responsible for what is said or done.
Luckily the EU is making some laws so that AI needs to be explainable, a reasoning needs to be easily extracted from some sort of easy UI. I don't know the specifics, though. Look up "explainable AI" (XAI) if you wanna know more.
You don't need to hold the AI accountable, you hold the party that created the AI responsible. If an engineering firm makes a mistake and manufactures faulty brake pads, they are held responsible. Likewise, if a tech company makes a mistake and produces a faulty AI, they can be held responsible.
We had a robo dept collection issue in Australia that caused people who couldn’t get any assistance with dept that wasn’t theirs to choose not to live anymore. These things are no joke when there is no avenue for people to take when mistakes are made.
Without regulation, one day you won't even get to talk to a human to try and resolve this. You'll call in and be forced to talk to the customer service AI. And when you file for arbitration your case will be decided by the arbitrator AI.
@@bananascoaster1243Not to mention the fact that people forget,that AI is made biased humans. I wouldn't be surprised if insurance companies come purposely tell the AI,to deny as much as possible.
Ngl I'd be very curious to see what kind of AI tricks people come up to circumvent some of it's preset behavior (if that still even is possible at that point)
This year saw changes to both ICD (diagnosis) and CPT (procedure) codes. If the insurance's AI program uses the updated codes, but the provider's EMR/EHR (computer records program) doesn't, then the insurance can pass this off on the provider. This is one of the many reasons why there are increased administrative positions in healthcare - dealing with the insurance companies' bs.
Good luck. Insurance companies keep hundreds of lawyers on retainer and pay millions in “lobbying” politicians to keep their grip on you. Try and sue them and you’ll lose everything trying to pay your own lawyers before they even set a court date
@@socialdeviant13what codes would have changed in the annual physical? Changes in the orthopedic world have been rather minimal, but maybe in other areas of the medical field they changed more?
But that’s the rub isn’t it. If she could afford to have a long drawn out, court battles with a multi billion dollar corporation, she probably wouldn’t need insurance. And these companies know it and most insurance companies the contract says your claim will be “ assessed” not paid they will drag it out for years mean while she’s without care.
And that's when you dispute the claim, request a review by a *HUMAN* doctor, and seriously consider filing an insurance complaint against them and changing plans to a *different* company.
sue them. sue the fuck out of them. they are denying you something you are entitled for and can't explain why, it's literally as good as them saying "just because". sue them for breach of contract.
I mean, if you want to buy your lawyer a new suit and possibly a new car.... sure, sue. BUT hold your horses, 1) contact the company and ask for an appeal or how to file an appeal 2) contact your state's DOI or 3) contact the BBB Once you've filed an appeal, had a DOI investigation, and had a BBB investigation, THEN and ONLY then do I suggest retaining a lawyer - but honestly if it's still denied, I'm sure you'll only be buying your lawyer a nice handbag. I work in the legal department of a health insurance company, I try my best to keep it out of court and I genuinely pity the idiots that pay their lawyer's a nasty chunk of change to go to court just get slapped with court fees and a reason to drink. Do. Not. Sue. An. Insurance. Company. Their contracts are written by absolute professionals, and I've never, in almost 10 years, seen an insured win in court. It's heartbreaking to see someone going through something tragic like a serious health event, so hell bent on wasting resources and time on crappy advice from someone who's laughing all the way to the bank.
@@bethannybiscuits Dude, they denied a claim for a PHYSICAL. Are you going to ask her to do all this shit every time she makes any claim? Insurance companies should APPROVE all preventative and diagnostic care because it is in THEIR OWN INTEREST to detect health issues early.
@stepheneickhoff4953 Yes. It's 1000 times more practical than suing. It was a simple bot mistake, and usually calling customer service or having it refiled fixes it. If that doesn't work, file an appeal. Or you can follow the sue them method 😉 idfc I get paid regardless. Stupid is as stupid does. I can assure you bot mistakes, like this, are so rare even the government is okay with it. States run audits on denied claims constantly because slapping an insurance company with a hefty fine in this economy is music to their ears. States love nothing better than to find someone slipping, it's easy money.
This is exactly what I said too. Well not "exactly", but yes it is absolutely a breach of contract and the insurance company needs to be held accountable.
The first line of defense for insurers is to automatically deny or delay. They are literally playing in your face; it's doing exactly what it was trained to do. It'll "change its mind" @ the 2nd or 3rd request.
Trouble is, she’s talking to a person who doesn’t have the know, the authority, and is under contract to complete calls. So they just don’t know. And can’t help. It’s at the ‘get me a manager/ceo stage now. And emails and letters to local politicians. The first responder just can not help
While I agree insurance companies are evil, they don't just deny every claim for no reason. It could be a very simple, valid reason for a denial. Such as she didn't wait 1 exact year from her last physical, her doctor recently went out of network and she has no out of network coverage, or the claim was submitted with an invalid/incorrect CPT code or missing necessary information. The only thing we can see here is a date, and the word "denied." Nothing else. This is likely from her claims list on her insurance app/portal, and she needs to locate the EOB which they should provide a link to via PDF etc. Esp if she's opted into electronic communication. The rep she talked to sounded like they were poorly trained. The wording made no sense, and who knows if that's exactly what was really said. As the above commenter said, she should try asking for a supervisor, or see if they have a process to allow the claim to be reviewed. Worst case she can file an appeal. Computers do a large amount of the claims processing these days, but sometimes mistakes are made. This is why it's good practice to educate yourself how insurance works, what coverage you have, and keep an eye on your claims so you can bring things to attention when you notice anything that doesn't seem right.
So I'm pretty sure the AI they're talking about isn't being trained to "approve correct claims," it's being trained to save the Carrier money. And to do that it's likely looking for claims or clients that have low activity on their plans so they're the least likely to notice a claim get denied (whereas a big procedure is likely going to be heavily scrutinized). Cigna got in quite a lot of trouble because they had a policy where they literally would DENY _NEARLY_ EVERY CLAIM and only those who called to fight for it would be approved (if it was supposed to be approved in the first place). Cigna did that so that they could lower the premiums to look good upfront on the market (cause insurance is already expensive, who wouldn't want to save $50 bucks a month) and once you signed up you'd be stuck there for at least a year without a qualifying life event. To top it off a yearly "subscription" is easily forgotten, so people would usually forget how much they hate their insurance and not change at open enrollment (November through March) so the Carrier would just keep passively collecting premiums without paying as many claims. Similar to Medicare Disability, all claims no matter how valid are denied first, then you fight for it and _THEN_ you _MIGHT_ be approved. They do that for a very similar reason: the budget is limited and they don't want to "waste" it, so they make you jump through hoops to get anything (which frankly probably costs more money than if they just did things properly). This AI is likely similar to that, except it's likely doing it more suitably so that it doesn't blow up in the Carrier's faces...again (which it will likely do eventually, just take longer).
Soooo... While I appreciate the info, I'm really taking away this: there's a way to use an insurance company's policies in a way that costs them more than if they had not pulled the bs of denying a claim in the first place...? If so, this should become extremely common knowledge.
@@llewane correct. Using your policy _at all_ costs them more, especially the free benefits guaranteed by the Affordable Care Act, as most people actually forget to get their free preventative physical every year. The other thing that costs them more is having clients that are not over insured; if you're young, with low risk life style, and little to no preexisting conditions there's frankly no reason to pay for a Platinum plan other than to give the carrier more pocket money as you are highly unlikely to use enough of your benefits to justify the premiums (unless you have planned events like pregnancy or repair surgeries). High deductible Gold or low deductible Silver is usually more than enough and can save quite a lot in premium. The other way to waste the carriers money is to call in to verify every claim is processed as they have to pay people to answer the phone/questions.
A computer can never be held accountable, therefore a computer must never make a management decision. IBM had that on a slide for training computer engineers in *1979*. How far we've fallen.
I’ve had this happen to me too. Multiple claims in a row, all denied for being out of network when their own provider database says that they are in network
I wonder if lawsuits will start coming out of scenarios like this. They don't have an arguable reason for the denial apart from the determination from a piece of software. And if the software is denying straight forward and routine situations where there hasn't been a problem in the past with claims processed by people, then it could be argued that the software is set up to deny standard healthcare support. What a frustrating situation, I hope yours gets resolved!
Cpmpanies use ia and fire employees to save money > companies get sued > companies lose a ton of money > companies realize it was actually cheaper to have employees
In my experience as someone with a congenital heart condition who has had to deal with insurance companies my whole life. The more persistent you are about an issue you're having, the more likely that issue will be dealt with to your liking. I'm a robot when it comes to needing to call once a day every day to see where/how/time frame of the process my claim is in until it's resolved. If it's not resolved to my liking and I have proof that a claim is being mishandled, I make sure I escalate my complaint so that it's recorded in their system. This is actually a lot more effective than people realize (depending on the complaint, I assume). If you have valid issues and your policy isn't fulfilling their end of the bargain, then talking to a lawyer is always an option as well. Most lawyers provide free consultations. I've only had to do this once and actually won my lawsuit. All that being said, you have to be your own advocate because no one else gives a shit about you in this life. Especially not insurance corporations or government programs that focus on healthcare (the government doesn't give a shit about you either, but that's a different topic).
Exactly - while not all appeals are approved, only a very small percentage of people even bother appealing. Insurance companies like this. You are right, you have to be your own advocate. I work in insurance and I have a feeling she spoke to someone new or someone who didn't know what they were talking about. When I call customer service, sometimes I just call back and get a different person and they're able to resolve my issue with no problem. Or ask for a supervisor. Sometimes that won't help if the company has upheld to the guidelines in your contract. But in the end, you have options. Use them.
YES. You are so right. It can be a full time job just managing all the calls to coordinate (and argue) these things for yourself. I'm sorry that your life has these struggles, too, but you explained it so eloquently, and I hope it will help others who are struggling. ☀️
I feel a swell of gratitude that I was born in Canada every time I hear about the U.S. healthcare horrorshow and the rentseeking insurance companies that lobby to perpetuate it
I learned this the hard way. I had an injury last year, and had to push for certain tests to be done that were valid and would help put me at ease due to the symptoms I was feeling. The scans showed I was improving and I stopped stressing as much because it proved to me there were not more complications. I was sad having to push and push for tests (brain scan, btw…not something to put off after certain injuries), and honestly it was emotional because it also showed me how people with harder conditions have to push.
Okay so, I actually used to work claims at a health insurance company about a year ago. It worked for them is most claims would run through the AI system. I think they said about 90% of claims. Then, any claims the system couldn't run would come to the claim's specialist (my job) and we'd run those. But, often if someone made a complaint about how their claim was resolved, it would often go to a claim specialist and they would manually check it. I have a feeling that "the AI sometimes changes it's mind" actually means someone is going to manually check it. I can't say that for sure but it makes the most sense to me
that’s kinda sketchy how they wanna double down on the ai thing like it won’t make us feel better to know a actual human a entity that doesn’t just exist on a computer is reviewing our shit
It's not ai, it's just an algorithm. It used a combination of icd codes and provider codes to see who is covered by whom. If the situation isn't accounted for, it usually fails to a user. In the OP instance, her provider and the insurance provider probably don't have termed contracts anymore.
My guess they are probably just denying all the claims they get hoping to save money on the people who don't want to bother appealing. Probably just using ai as an excuse.
The problem with that approach is that people with executive planning dysfunction won't be appealing either, not that they don't want to, it's much harder for them. And since executive planning dysfunction is an actual disability (like ADHD for example) that can be thrown into disability discrimination if they don't have legit reason to deny/delay.
Unlikely. AI halucinates. It will just make shit up. So in that run of the AI it just decided that "annual physical" was a miss spelling of "Anal Physical" and "anal physicals" wouldn't be covered under most circumstances. (this would be an example of AI halucinations, no clue if this is what the system actually did)
@@anythingbut...even though that sounds nice do you think the people who have that would actually be able to get a lawyer to sue if they can't even look into this?
@@TheMangoMovementi have adhd and i got laid off and never got around to having an attorney look at my severance package. Iv literally never had to contact an attorney for anything and i can’t describe the mental health issues that came with freaking out about finding a “good” attorney or worrying that they would tell me not to sign it, because i was also having to deal with the horrible unemployment system and well long story short i sold my soul and signed away my basic rights for $2000. Fml
As a former utilization review nurse it is 100% well that the insurance company must tell you in writing why they denied your medical claim. Also annual Wellness visits are covered 100% and in every state you have three levels of appeal and in some states you have a fourth level of appeal for claims that are denied and one of them must be a peer review buy a physician
"Lady, we asked Henry the octopus if you were getting medical attention. He said no. Come on the full moon at 5. We feed him, he can be more accommodating."
As a medical professional, nothing is more infuriating than seeing the health insurance that patients pay for act as a LITERAL BARRIER to appropriate care. Furthermore, AI should only be implemented into healthcare with great caution. This is the worst of the worst right here (edited for readability)
In CA, a 2024 proposed law would require insurance companies have a licensed physician supervise any AI-based decisions that approve, modify, or deny a patient’s care. So this is a real issue and hopefully laws can be enacted to protect patient care.
there should be no ai based descisions at all, never. cause if a physician supervises it, he can do it to begin with! Stop this shit already, its basically a racist anti-human black box. it learns from us, it learns SHIT then.
@@lucasm4299 Honestly what you need is a national healthcare system paid for through taxes that actually takes care of your citizens. Cut out the middleman BS.
@@BiancaNightshade SB 1120 (Becker) if you are interested in reading the bill: leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202320240SB1120 PS. That’s totally awesome you want to write your legislators. It’s so important to have our voices heard by those making the decisions. I work in healthcare advocacy so thank you!
Sounds to me like the AI is doing exactly what it was designed for: to deny as many claims as possible and give the company an excuse to not pay for your medical bills.
They need to immediately contact their states Insurance commissioner who oversees these companies to ensure proper application. More importantly, when our claims department received an inquiry from this office, we were required to respond expediently (usually 30 days) or face repercussions.
Thank you! Yes, she needs to look for the explanation of benefits. If it's not available on their app then it may be on the insurance companies portal if you log into the desktop site. But somewhere in the claims section should be an option to view the EOB. This is where the denial information is listed. I think she just spoke to a customer service rep who was new or poorly trained and didn't know how to access the information (or explain it)
Gotta love a field where objective proof and repeatable results are so important like healthcare getting segments replaced with the Vibes Box that's basically like asking a magic 8 ball
Because the AI was trained on their claims reps. Sometimes claims reps will simply deny 1 out of 5 claims, knowing that only 1/3 of policy holders will appeal the denial. Making it very profitable to randomly deny claims.
Sounds like a solid theroy to me. Especially if they hit a person with that piss poor excuse. It is sad as the thing is... You can always fight the claim. They don't always make it easy, but you always can. I would like to hear more of the story. I hope she didn't hang up, and leave it there. I am terminally ill, but when I was at my worse? Dealing with the bureaucracy for insurance and doctor offices was by far the worse part. Endless forms, letters, and phone calls. It was so frustrating being under 80 pounds, just being told Id be dead soon in a hospital on the phone for hours. Just figuring out what I need to do to get actual help.
Lawsuit. Have them explain why the AI finds it necessary and logical to breach your contract with the company and why the company finds it necessary to cover your medical needs with a chatbot.
"We can't tell you why your claim was denied" is insurance language for "we know exactly why it got denied but if we tell you why we'll have to pay you so we'll just keep silent so we can pocket the money"
I feel like a law should be made to make it illegal to use AI for any purpose other than entertainment
День назад
As a someone with a software engineering background, this AI model wasn't necessary. There are a finite number of ways to code medical claims (even if it's a big number of ways), and they have very limited responses (Approve, Deny, and in special cases, something like Escalated for Additional Review). I can understand using it to screen special cases that would normally require human review, but that should not include basic healthcare like an annual physical.
AI does not have a medical license and as such has no authority to withold treatment. This company is having an AI practice medicine without a license. File a complaint with the state AG and the medical board.
Legally speaking, the AI determined that the procedure did not comply with the insurance policy so that the claim was non-payable. That's different from practicing medicine.
@@sigmascrub that is effectively the exact same thing as withholding treatment, if you have to pay for the treatment. You shouldn't have to pay twice just to get an ANNUAL physical exam, else what is the health insurance for. I'm glad we have socialised healthcare.
This is absurd, especially if they can’t change it manually! A lot of the times it seems like AI is used as a excuse - we didn’t do what we wanted to do anyway, the robot did the crimes
If any program (AI or not) regularly "changes its mind", hire a new programmer. Reliability is important for any program but for Healthcare it is Required.
I work in P&C insurance. When we deny a claim we have to site the part of the contract we used to deny. No idea if health insurance has to do that, not all regulations apply to all types of insurance. Ask for a denial letter in writing and see what they say. Ask them to cite the part of the contract they used to deny. Also, yes 100% file a complaint with your DOI. The DOI will ask the company why they denied the claim. The company will be legally required to respond to them. If they get enough complaints they may investigate the company.
Insurance companies should be taken to court in a class action suite over their use of AI, because if a Fkn doctor has prescribed a medication or refers you to a specialist, than the insurance company SHOULD NOT be able to tell you that "oh no you dont actually need that". The over reliance on AI has become a completely scummy practice and insurance companies specifically need to be made an example of over their garbage.
If you called your insurance provider and this was your answer, “I would like to speak to your supervisor for a more complete answer” is in order. If that is denied, ask your phone provider for the call log or have the recording subpoenaed (they record it hence the little message in the beginning about quality control) and let the judge and/or jury hear what they had to say.
I love the enthusiasm but unfortunately that is not how it works. The majority of the times not even the supervisors can tell you why the database denied the claim because it does not even tell them why. A lot of times it's something super small. All you can do is resubmit the claim and try again until it works xD
The insurance company would be the one to pull and review call logs, and honestly supervisors don't usually have the magic fix-it button. Suggestion would be to have a claim resubmitted and sent for manual review
@@plopstopanddropp I found out this video was totally made up by another commenter. Apparently this lady’s whole personality is making stories up about AI because she’s afraid of it or something. So my question to you is, why’d you totally make up a response? How many times have you dealt with insurance companies? Every time I ask for a supervisor, what I want done gets done, or I get an adequate response that is backed by their terms of service.
Two things occur. Firstly this AI is hot… in a bad way. AI has to use some stochastic methods. However the more autonomy you give it, the hotter it gets. They call this entropy for a good reason. AI as a black box process is incredibly…INCREDIBLY dangerous. If no one knows why AI did something, no one can reverse engineer the process, and it makes weird and unwanted calls like this. Well, if a human did this you would call them with untrustworthy or unreliable. And that’s not a good thing.
If it was a human, they’d be fired. Since ai is comparatively cheaper than hiring people to do the job, it’ll stay and stuff will keep getting worse until someday new regulations will be instated.
Bruh. You should look up stochastic. AI is not a black box process. Of course you can reverse engineer the process. There are so many free AI courses you can take that show how to build them from scratch. These are nonsense words.
@anaryl what are you talking about, ai is literally a black box. We cannot see that direct processes that led to its outcome. We can only retrain and guess, it can make up a reason for why it did a thing but that's useless. Ai is dumb as rocks and shouldn't be used in ANY medical or law applications. Use it for unimportant shit
@@anarylYou can have deterministic and stochastic models, most AI is stochastic because it makes use of randomness when it's learning. And a major problem with AI is the interpretability of it, so yes a lot of AI is black box. It's a huge area of research because without interpretability and being able to trace back decisions that could be understood by a human problems like the one in this video occur. I would keep learning pal!
@@anaryl You’re the one who has no clue what you’re talking about. 😂 “Stochastic” does not mean we can get a model’s rationale for an output. It actually means the exact opposite of that 😂 Don’t talk down to people about stuff you know nothing about. The insurance AI is not chat/text based so we physically can’t ask them to walk through their rationale. Thus you in your infinite wisdom should know that the only other option for us to interpret it’s rationale is to look at the exact neurons that were fired during execution and then somehow gain some insights from that. That’s just as hard as looking at a human brain’s neuron firing and then trying to predict their exact thought process. (it’s impossible currently) You’re welcome, I hope you learned more than just about AI today: humility.
When computers first came in re billing & record keeping (for anything commercial) the common response was: "That's what our computer says". Followed by a refusal to check if the programming/information going in was correct or not. This seems like an updated version of that.
I work in AI. Explainability and Transparency are important aspects of a 'good' AI. If the company can't explain why their AI made decisions, they shouldn't use it. You are right.
In recent months calling dozens of lawyers, not one has said they can take the claim/suit because they know I can't win. And I don't have money to just pay them hourly to take case anyways after being told it's not going to win or even if the ly told me I can win, I still don't have money for lawyer hourly rate.
This happened to me with my son. The insurance paid for his 35k surgery but denied a follow up X-ray because of some questions about liability... I said he tripped and fell on the ground and broke his elbow. There's no one to sue and you already paid the 35k why deny the $100 x-ray? They just pushed it through again and it was all good
No, some use a stochastic input. The issue is that this is a classification algorithm essentially which has no reason to be non-deterministic unless it's accounting for other factors. Like say how much the insurance company has paid out already.
@@danielmorton9956 I expect they did so because the insurance company can train an AI on their entire claims database but there's no similar way to automatically generate a deterministic program. They would need to pay many expensive programmers to manually convert encyclopedic policies and procedures into piles of if statements.
@tylerpeterson4726 The programs are inherently deterministic unless built specifically to not be. But Here's the real catch- insurance has been using AI for a long time, before machine learning. You don't need to program in if statements. Decision trees literally are that automated. However, before ML methods they did use expert systems which are similar to if statements, but essentially all the procedures are already "coded". This makes for easy one hot encoding. The factors taken in come from chains of billed codes and patient data. There is no reason for, say, a generative algorithm. These are less deterministic.
To me it sounds like the staff member doesn't understand AI or is trying to be vague. If, for example, it was denied because of a clerical error by the doctor's office then the insurance company must notify the office in writing. The doctor's office should then be correcting the claim and resubmitting. Sometimes a bill goes out in the meantime. Then the AI "changes it's mind."
That's not really AI in this case, this is a common insurance practice to "accidentally" deny valid claims in the hopes that people won't contest the denial. It forces you to waste time to get the claim filled but they make X more dollars from the poor folks who don't contest it. Not surprising that they would hide behind the skirts of AI to pull the same scam.
i think it’s important to note that AI itself is not exactly the issue here. machine learning models at this point are more than competent enough to process almost all claims effectively. the issue is the COMPANY configuring these models/AI to unethically deny claims to save money. AI can be amazing or horrific depending on its user.
I love that in the EU, if an automated decision is made about you REGARDING ANYTHING, you have a right for a human review. It’s a less known part of the famous GDPR (article 22). You go girl!
I work for a Health Insurance TPA and this would be my nightmare. We re not using ai to process claims and I genuinely hope we don’t. We still use humans for CS as I am one. Still have automated system but you can speak to a human. I mean if the ai made a mistake then why not have a provision in place to have claim sent back to be adjusted correctly by a human?? It just smacks of incompetence. Not to mention I can imagine they are creating more work for themselves because more people are calling, emailing submitting Inquiries via their portal regarding claims being denied/incorrectly processed via their AI.
If payment is her responsibility, legally she needs to be able to appeal and they need to provide her with all the information to do so. Which means it's on the doctor's office to appeal. She's all set. I'm guessing the AI "changing it's mind" means the doctor's office made a correction and resubmitted, and the service rep is either being intentionally vague or not trained with the AI.
Engineer and ex-NIPR licensed healthcare producer here. Document everything and talk to a lawyer that specializes in health insurance class action lawsuits. This may be illegal in a way that would make class action lawyers drool. There's laws around unsupervised claim denials, especially denials of covered services without any stated justification.
I recently read the article "I will f***ing piledrive you if you mention AI again" and it FILLED ME WITH VINDICATION also the "sometimes it changes its mind" is just W I L D
As someone whose mom worked in insurance, every ins company has a Grievances and Appeals process/department. Use it. That's some bullshit, and appeals is how you fix it.
As a sci-fi nerd, the thing that irks me the most and "AI" is that what actually, currently exists is so far from what the average person thinks when the hear the term that it feels misleading to even use it. There is no "mind" in your computer, Becky, it's just a calculator that can use words as well as numbers! *LAAAAAAAME!*
The majority of AI "tools" I have seen out there are simply just machine learning algorithms. Most of them don't even meet the definition of AI. It makes me cringe every time I see it..
@@theundone777 problem is the definition of ai is about as broad as the definition of what is or isn’t ‘living’ even some of the simplest pieces of code count as ai because it isn’t referring to any particular level of intelligence beyond that of bacteria
I'VE VEEN SAYING THIS! LLMs and newer "AI" built off those are just calculators for words. But because humans have literally never experienced anything that could use language without it being connected to a conscious living creature, our silly little ape brains have a hard time not ascribing some higher order consciousness to these language models.
Exactly. It's like packaging for snake oil. It's why there are so many assholes who look at the unethical regurgitation machine that can't create anything without scraping artists' work and say "humans learn the exact same way!" 🤪