How will AI impact diagnostic radiology in the next 50 years?

Technology Forum
1

Can someone share their honest thoughts on how AI might change the future of diagnostic radiology over the next 50 years?

2

Welcome to this forum

A Few Guidelines for Posting and Discussion

Please follow these rules for current and future posts:

  • Posts must be over 100 characters – more detail helps.
  • You might find your question already answered. Please search first if no one is responding.
    • There are a lot of posts on AI and job loss, just a heads-up!
  • Feel free to discuss both the positives and negatives of AI, but always be respectful.
  • Provide links to support your arguments.
  • Don’t ask dumb questions, unless you want to discuss AI bringing about the end times – that’s not happening.
Let us know if you have any questions or comments.

I am a bot, and this action was performed automatically. Please contact the moderators of this forum if you need help.

3

You’ll see a lot of opinions based on wild guesses about AGI here, but that’s not what experts are really worried about when it comes to radiology automation. People have been concerned about radiology being automated for years, even before people started worrying about job losses with LLMs.

4

Ah, the AI vs. Radiologist debate… it’s like Netflix vs. the cinema, but with more DICOM files and fewer popcorn buckets.

In my opinion, this should (but probably won’t) go from ‘helpful assistant’ to ‘invaluable colleague’ and then to ‘colleague who does all the boring stuff but still can’t run the Friday MDT.’

I’d be happy for this AI buddy to handle the on-calls, but I’d still want to do the daytime work for my (hey, pretty #GMC) appraisal.

So maybe diagnostic radiology will be like Iron Man with Jarvis: the AI does most of the work, but there’s still a human making the final decisions.

5

I’m a radiology tech and software developer. AI has already made a big impact on radiology interpretation, starting with computer-aided diagnosis. I predict that in 10 to 20 years, the radiologist’s role might just be to rubber stamp AI results most of the time.

6

@Indigo
I hate to break it to you, but it’s basically already happening. It’s more of a firewall issue, but you can bypass it with large models if you know how. The accuracy is already really high.