AI creates images of what people see from just an fMRI

  • Thread starter Thread starter BWV
  • Start date Start date
  • Tags Tags
    Ai Images
AI Thread Summary
Recent advancements in AI technology allow for the generation of images based on brain scans, showcasing a direct comparison between what test subjects see and the algorithm's interpretation. The discussion highlights both excitement and concern regarding the implications of such technology, particularly in understanding human thoughts. A humorous observation was made about a generated image featuring a teddy bear with a Hitler moustache, prompting questions about the subjects' mental states. There is speculation on the potential of NLP algorithms to further decode thoughts, suggesting a future where understanding human cognition could become more accessible. The conversation reflects a mix of fascination and caution regarding the ethical implications of these developments in AI.
BWV
Messages
1,571
Reaction score
1,925
Here is the image viewed by the test subjects vs the image generated by an algorithm interpreting a brain scan of the subject

1677869348644.jpeg


https://sites.google.com/view/stablediffusion-with-brain/
 
  • Wow
  • Like
Likes DennisN, berkeman and Borg
Physics news on Phys.org
What a time to be alive. Thanks to the advances in technology we can finally test if it's really true that when all you have is a hammer, everything looks like a nail.
 
  • Like
  • Haha
Likes Jarvis323, DennisN, Bystander and 2 others
Swamp Thing said:
Why does the teddy bear have a Hitler moustache?
Have to ask the test subject, although perhaps weeks of therapy might be required to get the answer from them …
 
I find this exciting and scary at the same time. If you use an NLP algorithm instead, it isn't a big leap to know what someone is thinking about.
 
Borg said:
I find this exciting and scary at the same time. If you use an NLP algorithm instead, it isn't a big leap to know what someone is thinking about.

Careful, comrade.
 
Swamp Thing said:
Careful, comrade.
Not what I was thinking. :wink:
 
Borg said:
I find this exciting and scary at the same time. If you use an NLP algorithm instead, it isn't a big leap to know what someone is thinking about.
Wouldn't an NLP version be something more like subvocal recognition (which has been around for decades)?
 
Back
Top