Using AI for Your Disability Case? | The Disability Benefits Podcast #41

In this episode, we discuss the potential dangers when using AI for your Social Security disability case.

Click Here to See if you Qualify For Benefits

Disability Consulting, LLC
(877) 204-8159

Disclaimer: We are a non-attorney firm. Our company specializes in representation for Social Security disability claims and is not a law office. The information provided in this video is for general informational purposes only and does not constitute legal advice. For legal guidance specific to your situation, please consult a licensed attorney.

Transcript

SPEAKER_01 0:00

All right. Well hop into it. Tell us about AI. All right. So we now live in an age where everybody is using AI, and everybody's using AI all the time. Like I just told you, you know, if I'm at a ball game and I'm like, man, you know, he's he's he's got 25 home runs in June already. You know, what's the most home runs that were we're hitting, you know, 1965, the year I was born? Um and you can type that into Chat GPT or type it into Claude or type it into the AI of your choice, and you're gonna get that factoid. And most of the time it's gonna be right because it's just a factoid. Um, unfortunately, what we have encountered, not so much, but it's gonna become more and more a problem, are people typing into Chat GPT? I'm applying for Social Security Disability, and they've sent me an ADL form. What should I write? What should I write on an ADF? Chat GPT, he's gonna tell you. He's gonna tell you exactly what you need to write on uh uh uh activities, a daily living form. And God help you with your claim if you write that kind of stuff, because it's gonna be wrong, wrong, wrong. It's gonna be based on the other three, five, ten thousand people who have asked that question, or maybe have put in something that the AI is now pulling from, whether it be Chat GPT or Claude or Plexity or any any AI program that you're using. Um I would highly encourage you not to do those kinds of it's personal me, you know, those kinds of questions and ask a AI to help you with that because it is much more credible for somebody at Social Security to see your chicken scratch and your handwriting is terrible on this form, but doggone it, it's yours. You actually answered that question and it's yours. It's it's not something that, you know, artificial intelligence, which it's definitely artificial and it's rarely intelligent, in my humble opinion, um, that they're trying to, you know, trying to give you the answer to win your case. Okay, yeah, you want the answer and you want to win your case, but you certainly don't want a manufactured answer. Because one, it's not true if it's manufactured. You want it to be true. In fact, we've talked about this before. If you fill out these forms, they become sworn testimony later on. You know, whether the judge, if you hit if you make it that far to a hearing, the judge might say, Okay, well, I see that you know, you said that you you know you can't take care of these uh three horses that you've had, you know, since you were a teenager. Um you went into the exam and the guy said that you smelled like a barn, and you said, Well, yeah, I've been taking care of my horses. Which one is it? You know? Um it would make more sense if you're just telling the truth. And so to ask AI, what should I write, is horrible. Yeah. Same thing. Let's go a little bit further. You're obviously you're on your own because if you have a rep, they would probably tell you you steer clear of AI with these kinds of questions, they're gonna mess you up. But let's say you don't. You don't have a rep, you're all by yourself, and you've been denied twice, and you're at the hearing level, and boom, you get that letter in the mail. Your hearing has been scheduled for August 15th at 3 o'clock. Okay, at August, August 15th at 3 o'clock, I gotta be for this hearing. You go on to your favorite AI program. You know, what do I tell a judge in a disability hearing? It's gonna tell you what to tell a judge in a disability hearing, especially if you say something as foolish as what do I tell a judge in a disability hearing if I want to win my claim? Yep. Oh my goodness. There are no ethic parameters or ethical parameters. You just asked it, what do I need to say to win my case? You didn't ask it, you know, to help you truthfully express what's going on in my life. You didn't ask it that. You asked it what can you, and it's going to give you an answer. And once again, God help you when you get that answer. You go into the judge, and um, you know, let's say that because you know it's a phone hearing, and you know that the judge can't see you, you've either written down what AI has told you, or you have your phone or your iPad there, and you know, well, judge, I'm disabled because and you start reading. Let me tell you, these judges are highly trained individuals, and you do that kind of nonsense, they're gonna come right out and ask you, are you reading something? And you better tell them the truth. And when you tell them the truth, yeah. What are you reading? Well, I asked, you know, Chat GPT to tell me how I can win my disability hearing, and it told me that I needed a word things like that. All right, well, you need to put that away, and I'm gonna ask you some questions, and we're gonna go through this the way it's supposed to go. It will not work well for you. Um, and we are in that age that we really count on that, and like if it's a factoid, um, it's usually the information is usually correct. But when it's asking you, how do I go about this? Or, you know, how should I word this or how should I do this, there are no guideposts that it has. Nothing's gonna stop it from saying something that's absolutely ridiculous, not just ridiculous, but untrue. And if you start telling Social Security things that are not true, just to win your disability case, they have a word for that. And the word is fraud, and it's you want to avoid that, um, it will follow you forever. So, you know, my little snippet is you know, we live now in the age of AI. I use it. I love checking out factoids, I love um putting in uh travel information. You know, how long is the flight from this to this? You know, if I land at this airport, how far is the drive from this? I love doing that and I love having it at my fingertips. And there's some other things, just you know, silly factoid things, you know. You know, when was the first uh light fixture, you know, in America? You know, when when did the first bulb light up on electricity instead of gas or whatever? You know, I that that those are kind of fun to look at. But to use it in something so personal and something so subjective as disability, um, you're just flirting with disaster. I I I can't see, I I literally can't see any favorable result from you with using AI to help you fill out a form or help you words word something so you you feel better about it. If you got a fourth grade education and you word something like you've got a fourth grade grade education, fine. Yeah, nothing. What's wrong with that? That's the truth. You misspelled, you know, half the words. So what? You know, that's the truth. You get uh AI to polish something up for you or to put something in a way that you would never put it, you wouldn't even think of it that way, and you don't even recognize it as yours when you read it. It's the worst thing you can do. And once again, it will not go well for you. It just won't. It's it's gonna be recognized by if it's not recognized at the lower levels, which it probably will be, but it's definitely gonna be recognized by a judge. And the nice judges are gonna say, put away all that garbage and we're gonna do this the right way. And some of the not so nice judges might not say a word and might just deny you by in in the denial letter calling out everything that was a contradiction. Well, the medical record showed this, and he said this, and that wasn't true because the medical record showed this, the form that he filled out is definitely not by him. You get that kind of stuff, and it follows you. And that's the last thing you want. You don't want some some horrible mistake to follow you, and you can't get out from under it. So just don't make that mistake. And that's my you know, cautionary tale on AI.

SPEAKER_00 8:46

Yeah. You know, when I'm doing intakes, when I'm screening someone to see if they qualify, and also if I think that it's a case that we can win, so we want to take it. I'm very reluctant to uh say certain symptoms or diagnoses because I don't want to put words in their mouth. Sure. And one of the most common things is neuropathy with with diabetes. Now, neuropathy is not the same thing as numbness and tingling. Neuropathy is you've done testing and you have neuropathy. Right. And so what I don't want to say is I don't want this person to tell me they have diabetes, and I say, okay, and do you have neuropathy? And they're like, What's neuropathy? It's like, well, it's numbness. Yeah, I have neuropathy. And it's nowhere to be found in their medical. You're gonna put that down, and the worst thing I could do is like, okay, so you're diabetic, you have neuropathy, right? Okay, and so it's probably bad in your feet and your hands. So you're just probably just leading them along. Yeah, I'm just leading them along there. So I'm very hesitant to say stuff like that. I w I want to say, what what are the main symptoms that you experience from your diabetes? And then do you have any numbness or tingling in your hands or feet? And then a lot of times they'll say, Yeah, I have neuropathy. It's like, okay, you said it. Yeah, sorry. You I didn't say that, you said it because you're diagnosed with that. You've you've done the testing. Right. My main complaint with AI, and this has been the case for you know several years that I've been using it, is it's so agreeable that it's going to give you an answer no matter what. And if you ask it an objective question, it's gonna give you an answer and brand it as a objective answer. And if you ask it a subjective question, it's gonna give you an answer. And it's probably not gonna say all of these disclaimers and all of these, like, you know, discrepancy between it's just gonna give you an answer. And so you might type in, how do I I'm I have a hearing tomorrow, how do I win my case, what do I say? It might tell you about how you should talk about, you know, you can only walk or stand for 15 to 30 minutes. It's like, my diagnosis is depression.

SPEAKER_01 10:44

That's right.

SPEAKER_00 10:45

But you're gonna you're gonna you're gonna see that. And if you don't know anything about disability law and and maybe you're just desperate to win your case, like, okay, yeah, I'll I'll tell him I can't walk it stand. It's like, well, I don't see the judge is gonna look at that and say, like, why can't you walk or stand? The good ones will. Yeah, right.

SPEAKER_01 11:02

The bad ones will be making a note and saying, okay, this guy's just holding my leg and I'm gonna deny him because he's he's lying through his teeth.

SPEAKER_00 11:08

Yeah, it's like if you if you ask AI, if you go to whatever language model you use and you just say, What are the grid rules? That's fine. That's a that's a good thing to look up. It's like what is define the grid rules, what's the deal with having to be or being 50 and then your disability case gets easier? That's fine. That's totally fine. But you can't say I have degenerative disc disease, what are some symptoms that I can have that will win my case? Because it will tell you the symptoms. Oh, yeah, sure. And then if you use those, then like you said, that's that's fraud.

SPEAKER_01 11:38

Yeah. And you, you know, going back to you know, looking up certain things, there, you know, there are certain Social Security rulings, SSRs, that that are very common in a disability hearing, the grid rules, you know, if you're over 50, they're very common for physical cases. But I would even warn you against that because if if you're looking for factoid, why not type into Google SSA grid rules?

SPEAKER_00 12:11

And go to SSA.gov where they've got it.

SPEAKER_01 12:13

And there it is. Yeah, yeah. It will actually be the CFR that has the grid rules. And there it's a grid, by the way. So there are several grids. Um why not just do that? And you can read it. This is what this means. It's actually fairly um easy to understand. And I think that's a lot less dangerous than typing it into AI and um and letting them put their spin on it. Now, sometimes they'll just basically say, well, this is from you know SSA.gov or this the CFR that has the grid rules in it. Um okay, all right. But it's just it's dangerous um because if they give you that snippet of information over here, but you needed it right here. This is what you're really asking for, but they gave you this, and so you end up using this. And it's like you you needed this if you had asked about this to the source and not you know trying to get AI to clean it up or sweeten it or paint it a pretty picture or whatever. You know, if you needed that, you should have gotten that, and you you got something over here and it was pretty, and therefore you're gonna accept that and you're gonna use that. And likely half the time, whoever you're using that with, whether it be a judge or the lower levels, they're gonna be scratching their head. You know what? Why why are you telling me that or why are you presenting that? You know, that that really I don't understand why you're doing that. This is what is this what you mean? Oh, you know, it's so yeah, as you can tell, I haven't had the greatest experience, you know. I you know as of now, I have no problems with the A, the artificial, but I really do think that when you get out of just like an encyclopedic kind of uh function, that it just fails. You just need to be careful with it. It just fails.

SPEAKER_00 14:14

Yeah, it's like it's like so many other things, like using AI to code for you. The thing with getting AI to generate code when you don't speak the coding language is if there's a problem in the code, you can't, you don't you don't know the language. Right. So you're not gonna be able to tell. And so it might make this this code and then you put it on your website, and then something's broken on your website, or the application that you're trying to get AI to build is broken. You can't troubleshoot it. You don't you have no idea what you're looking at. So you're just gonna trust it and you're gonna copy and paste it in there.

SPEAKER_01 14:44

Or you're gonna call it on its mistakes, and then it's gonna agree with you.

SPEAKER_00 14:47

Yeah, then it's gonna give you a new mistake. Yeah, and just pr pretend like it fixed it. Yeah.

SPEAKER_01 14:52

Well, well you're agreeing. I just called you something really bad. You know, you're a useless piece of garbage. You know, I I asked for this and you gave me that. You're right. I apologize for that. It must be very frustrating. Here is you know what what I believe you were asking for, and it's more garbage.

SPEAKER_00 15:08

Yeah, it just doesn't change anything. Yeah, and likewise, if a non-disability representative puts in some kind of social security disability question, they're gonna get an answer. Maybe it's true, maybe it's false, maybe it's just obscure, they're not gonna know because they're not a disability representative. Right. If a coder who speaks, you know, who knows JavaScript and JSON and all of these coding languages, if they use these AI models to um make a string of code and it's there's an error, it's like, well, there's the error because they're a coder, they know what they're doing. And the same thing applies to you. If you used AI for this reason that you're advising claimants not to, you would be able to tell, well, that's not right because you're an accredited uh accredited disability representative. That's the whole point. It's like just keep in mind you're using using platforms that are not accredited, you know. You have SSA.gov, you have social security themselves, and then you have accredited disability representatives, people that are actually accredited to talk about what we're talking about and do what you do. Yeah. So using something like you know, Chat GPT or Perplexity or and we're not affiliated with with any of these companies.

SPEAKER_01 16:17

We don't love them or hate any of them.

SPEAKER_00 16:19

Yeah, or maybe maybe we hate some of them, but we won't name which ones that we do hate. Yeah, that's not a government tool, that's not a disability tool. That's just something that you're trying to force to work for you, and sometimes it does, but it's just you really gotta be careful with that kind of stuff.

SPEAKER_01 16:32

Yeah, you have to be careful. And um, you know, I think one day, maybe, but that day isn't today. And um when we were at uh a conference of the professional representatives organization that I belong to, this has been about two or three years ago, and the guy get up got up there and he he he had AI do something that was pretty impressive. And then he said, he declared, he said, you know, today is the dumbest AI will ever be because tomorrow it'll be smarter, and then down the road it'll be a lot smarter. And that's probably true, but we're still in a really dumb for what we do, for subjective, I need to tell you about my pain, I need to, I need to quantify sometimes the unquantifiable. I need to do that, I need to be able to talk to you as a human being. You know, how how bad does it hurt? It hurts so bad that sometimes I don't want to get out of bed. AI, can you put that in a in a better way? Nope. No, no, there's no way they can put anybody because as a human being, I know it's so bad you don't even want to get out of bed. Okay, all right. Why turn it over to AI and have them, you know, sanitize that, make it clearer and stuff? No, just say what you need to say. Um, and once again, it might down the road be something that can help um help in these areas, but I haven't seen it yet. And you know, and I'm you know, I've got my finger on the pulse of disability. I'm doing it for hours and hours and hours and hours every single day. So I haven't seen it yet. And I just caution um claimants, whether you're you're our claimants or whether you're out there, um, just be very, very careful. Um, you know, when when you're trying to use AI to express what you should just be able to express as a human being. Just express it, you know. Put it put it in a way. You can use a colloquialism, you know. You know, colloquialism, you can use a turn, a turn of speech, a figure of speech. You can do that. The judges are gonna understand that because they're human too. You don't need a computer to I want to fancy it up and I want to make sure it's pretty. Um, you know, I don't don't want it to seem like I'm some, you know, old country boy. Well, what are you? I'm an old country boy. Well, that's the way you are, then you know, that's the way you talk, that's the way you interact, that's the way you express. There's nothing wrong with that. It's an honest, uh, that's a very honest uh way that you uh present yourself. Um having AI, you know, clean clean things up or sanitize things, or it's just not gonna work. It's it's a hundred percent it's gonna accomplish the opposite of what you wanted it to accomplish. Yep.

SPEAKER_00 19:21

If you have a question about your disability case, drop it in the comment section below, and we will do our best to answer it in the next episode.