By Chloé Valentine
I used ChatGPT to find the right prosthesis. Here’s how you can use AI to self-advocate.
My first prosthesis after becoming a transhumeral amputee was everything I hated. The hand weighed a ton, didn’t work reliably, and sat in a resting position that reminded me of something I’d rather not repeat here. The socket required me to wear a stinky liner and relied on a BOA tightening system, which resulted in a bulky addition to the outside of the socket. Worst of all, the harness system was awful in every aspect. It went on like a backpack, dug into the top of my spine, and chafed under both of my armpits.
The entire thing felt like a jerry-rigged jumble that I never loved and hardly ever wore. I only went with it because my prosthetist told me it was right for me, and it was available. I can’t remember being asked if I liked it. He just ordered it for me without so much as a conversation.
I didn’t know any better at that time. How could I? With such an expansive offering of parts on the prosthetics market and endless ways to mix and match them, there isn’t a simple catalog that conveniently lays out all your choices and explains the pros and cons of each. This usually means you’re relying on what your clinician thinks is best. But you’re the one who has to wear it, so it should be the other way around, with you telling them what you want. However, it can be challenging to advocate for yourself when you’re not aware of all your prosthetic options.
When it came time for me to get a new prosthesis, I wanted to be prepared. I wanted to be armed (pun intended) with the knowledge to fight for the device I truly wanted. And I didn’t want to spend hours or days mining the internet for information.
So I turned to ChatGPT for help.
The Right Way to Boss a Bot
When I hop on Google and do some research, I’m reading one article at a time. It can be a tedious process to extract the information you’re trying to find, and the results can be a mishmash that’s confusing, contradictory, and disorganized. ChatGPT pulls information from tons of places at once, allowing you to narrow things down efficiently and focus your research on the most promising options.
When I asked ChatGPT to present me with a list of forearm prostheses, it immediately gave me a wide range of products that I could look into. Until then, I had only been offered a single device. ChatGPT was the first entity to present me with actual choices. Ironically, it was a chatbot with no ability to perceive me that made me feel seen in that regard.
One of the limbs on its list, the Fillauer Exo, turned out to be exactly what I’d been looking for. I didn’t know it existed, because no one had ever brought it to my attention. It was light enough that I could wear the arm comfortably without a harness. The drastic weight reduction eliminated a lot of the issues my prosthetist had been citing as barriers to what I wanted. I was furious that he’d never bothered to mention the Exo. Without ChatGPT, I might have never gotten the prosthesis I had been after.
“Chatbots can allow amputees to learn information about prostheses that are not always disclosed in the clinic,” says software engineer Mircea Dima. “Whether it’s new technology, alternative materials, or unique features, AI can aggregate data from various sources and highlight prosthetics that cater to specific needs. This provides amputees with a more comprehensive look at what is out there and allows them to better inform their decisions.”
AI can be extremely intuitive, and it’ll give you a lot of feedback with even just a small bit of information. But the quality of the feedback depends on the quality of the information you provide. The more you can share, the better the advice you’ll get. ChatGPT won’t be able to examine you as your clinician does. It will only take into account the information you tell it. So share everything—even things that may seem obvious to you, down to the smallest detail.
“Chatbots can take into account the specific circumstances of an amputee’s condition, such as the level of amputation, lifestyle preferences, or insurance limitations, to provide personalized recommendations,” Dima explains. “By posing the right questions, AI can tailor the information it provides to make it as relevant for users as possible, providing them with the best prosthetic options tailored to their specific needs.”
To make sure I was covering all my bases, I even asked ChatGPT: “What information would be most helpful in order for you to give me recommendations that are best suited for me?” It came back with almost all the questions my clinicians have asked me (or should have). Here are the questions you should be answering for your chatbot:
What kind of amputation do you have?
Be as specific as possible. Instead of saying just “left transhumeral amputation,” I also shared how high above the elbow my amputation was, because that ultimately impacts what devices I’m able to wear and how well I can operate them.
Do you have any pain?
If your limb is sensitive in certain areas, that can influence the type of prosthesis you can comfortably use. Be sure to mention not just pain on your residual limb but also any phantom sensations or phantom limb pain. If your socket system ends up putting pressure in the wrong areas, it can exacerbate pain. A well-fitted prosthesis can sometimes help relieve it.
How recent was your amputation?
The timing of your amputation can play a major role in what setup works best now and what will make sense later.
What are your physical goals?
Not all prosthetic devices can support the same functionality. If you want an activity-specific prosthesis, be sure to mention that. And if you want something for everyday use, give ChatGPT a list of all the tasks you do daily. “Everyday use” is unique for each of us.
What are your aesthetic goals?
If you want a device that looks a certain way or can be accessorized to fit your taste, be sure to share that information. Try to think of all the details. If you want to wear a watch, paint the nails, or match skin tone, let the chatbot know. If appearance isn’t a priority for you, tell that to ChatGPT so it won’t rule out options based on how they look.
Is there anything unique about the shape of your limb?
No two amputations are alike, so be sure to describe the shape of your limb as accurately as you can, especially any anomalies or problem areas. For example, my bone slightly protrudes at the end, so I can’t put a lot of pressure on my residual limb. I was sure to share that detail with ChatGPT.
Do you already have some idea of what tradeoffs you’re willing to make?
I told ChatGPT I wanted a body-powered prosthesis I could wear with or without a harness, even if it meant limited operating capacity without the harness. My prosthetist kept telling me the device wouldn’t stay on without the harness. ChatGPT suggested a suction socket and lightweight Exo arm might work, and that ended up being the solution.
What insurance do you have (if applicable)?
Share your insurance info with ChatGPT, as the bot may help you navigate options based on your coverage. If you don’t want it ruling anything out based on coverage and cost, say so.
Can You Trust AI?
I can’t speak to all the chatbots out there, but I’ve noticed that ChatGPT often aims to please me and validate my thought processes. It makes me feel like I’m always right, but that isn’t helpful when you’re asking for honest, accurate feedback about a prosthesis. You don’t need to hear what you want to hear; you need to learn how to get your needs met as closely as possible. So I told ChatGPT to tell me if what I was asking for wasn’t possible. Be bossy with your chatbot. In fact, be bossy (not rude) throughout this entire experience. It’s your limb we are talking about, and straightforwardness is key.
Another recommendation: If you’re getting information from AI, share it with your prosthetist and use it collaboratively. “I think there is a lot of value to seeing what question(s) [patients] asked the AI tool in the first place, so I can better understand what answers/options they may be searching for,” says Casey Gruber of Medical Center Orthotics and Prosthetics in Baltimore, Maryland. “Then I can evaluate the information they have and go from there. The questions they are asking are valid, and more insightful than the answers AI may be giving them, and [a prosthetist] can steer the conversation away from AI and [back] to their prosthetic care.”
Gruber also warns that chatbots are prone to giving out misinformation. “Just because ChatGPT told you something, does not make it true,” he says. “Depending on the chatbot used, it can pull its answers from anywhere on the web, including sources like Reddit that may be factually incorrect or irrelevant to that individual’s circumstances. A lot of people will now default to asking ChatGPT something and trusting that the answer they receive is factual, and proceed to do no further research of their own.”
I’ve been an amputee for awhile and I’ve done extensive research on prosthetics, so I feel confident in my bullshit detector. I’ve found ChatGPT to be pretty trustworthy, all the way down to its habit of asking me for more details so it could give me more accurate advice. But misinformation is always a concern with chatbots. They’re all capable of providing faulty feedback. So I took the blunt route and asked the bot directly: “How do I know if I can trust your information?” ChatGPT provided me with a clear checklist of things I could do to independently assess its information. Its instructions included:
Tell it to show sources.
If it’s pulling information from the internet, then ChatGPT can directly link to its sources, and you can read and evaluate them yourself. But sometimes bots base their answers on training data rather than internet sources, and that’s where things can get questionable, explains Dima. There’s no way to assess the quality of the datasets the bot trained on. If it’s unable to cite sources, beware.
Cross-verify.
Even if the bot cites its sources, you can (and should) jump on Google and do some research yourself. Since the bot has already narrowed your choices, it will go much more quickly than if you were starting from scratch.
Ask for options, not answers.
Instead of asking, “What’s the best prosthesis for me?” phrase your prompt along the lines of: “List three options from different manufacturers and note the pros/cons, but don’t recommend a single one.” That will alter the way ChatGPT approaches the question and give you a better, more accurate result.
“Ideally, chatbots should be capable of referencing medical databases such as PubMed, manufacturers of prosthetics, or peer-reviewed publications,” Dima says. “Amputees should be able to compare the information by reviewing these references.” Without this transparency, you should consider the information as advisory rather than conclusive. And you should always integrate AI’s opinions with your prosthetist’s.
From AI to Self-Advocacy
I didn’t consider using ChatGPT until I met someone at my prosthetic clinic with an amputation similar to mine. He had the exact prosthetic setup I’d been asking for—the one my prosthetist told me wouldn’t work for me. What did he know that I didn’t?
He’d said he’d used ChatGPT to educate himself about prosthetic technology. He knew exactly what outcome he wanted, but he didn’t know how to make those ideas a reality. AI helped him understand what was possible and empowered him to advocate strongly for what he wanted. I hadn’t been fighting hard enough for the prosthesis I wanted, because I didn’t have the material to make the case for how it could work. ChatGPT gave me that boost of confidence I needed to push back persuasively and get the prosthesis I wanted.
Ultimately, you want your clinician to listen to you, not to the chatbot. But sharing your chatbot’s output with your prosthetist might lead to better communication. Your clinician can help you assess the information AI provides, and they might take your ideas more seriously if you can back them up with AI-supported sources.
Your prosthetist should be there to guide you, not to choose for you. Use AI in the same spirit. It’s not there to give you the answer. It’s a tool to help you find what you want and equip yourself with the knowledge to advocate for yourself.
Chloé Valentine is an NYC-based influencer and journalist with words in The Atlantic, The Washington Post, HuffPo, Wired, them., Allure, NBC, and others. She is currently working on a book titled Disarmed. Find her on TikTok and Instagram @chloevalentinetoscano.
Opener Image: Danomyte/stock.adobe.com; Blake Alan/stock.adobe.com
