this post was submitted on 03 Apr 2026
0 points (NaN% liked)
Technology
84043 readers
7190 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
Devil's advocate...this is one of the few good opportunities for ML. If you train a model on a specific dataset with expert validation, this has the opportunity to save lives.
First, radiology isn't one thing; different radiologists with different expertise looking at the same imaging can see different things. Second, there are not enough radiologists; my wife is an ER doctor who only does overnights and her hospital network has a central radiology center that reviews all films from all the hospitals, and it's always backlogged and waiting on results impacts outcomes in a real way. Third, there are simply human limits to what we can visually perceive, take a look at this study: https://pmc.ncbi.nlm.nih.gov/articles/PMC3964612/
Radiological ML models could change healthcare. Imagine a world where part of your annual preventative care you go and get a full body CT. The ML model can compare your CT with references in your demographics AND your prior years, and find changes/issues before they're crises. That's simply not an amount of data analysis that could be done by an army of radiologists, and has the opportunity to spot things like tumors or organ swelling way earlier. I get that the late stage capitalism reality is "they'll use the data to farm money out of you" but from an actual technological standpoint, this could have real life-saving and improving implications for a lot of people and removes a huge bottleneck in healthcare.
Even if AI does the job of reading medical imaging extremely well, I’d still want a radiologist to double-check the scans.
I can't wait for this bubble to blow up in all their dumb faces.
For what it’s worth, “AI” in this context is probably not the content-stealing Generative AI that everyone is trying to cram everywhere it doesn’t belong. This is a much more legitimate application of a similar technology.
I’m not mad about the idea of AI in radiology because it’s a really good fit. A human radiologist can’t compare a hundred similar slices and cross-correlate possible anomalies, whereas AI can. This improves detection and outcomes and is exactly where medical technology is supposed to help.
That said, I don’t think we’ll replace radiologists across the board for a long time. This will be a very useful tool and will probably reduce the number of radiologists required and modify their roles significantly, but it’ll be more like how a single worker with editing software can do work that would have required a small team in the pre-digital days of film.
The replacing part is the problem. Using a local system to help is fine, but it still requires humans who know what they're doing and what they're looking at.
Sometimes, for example human + AI systems used to be better than either one in isolation, but chess AI improved so much that the human partner is actually not helping anymore
But chess is an isolated "system" with clear rules. Reality and especially medicine is so much more complicated.
Chess strategy is extremely complicated and probably will never be completely solved. It will be almost solved like checkers eventually when programs will just draw vs. each other or a white win is found
But we will never actually simulate all games since the number of chess games dwarfs the number of atoms in the universe. So in that sense we will never know what the "correct" move is outside of table base or mate situations. Medicine may actually be less complicated to a machine.
Bu the only benchmark should be "how good the humans are at a task" since you're not trying to be perfect. You only have to provide better results than the current system.