Sun-ha Hong
LightNews LightNews
banner
sunhahong.bsky.social
Sun-ha Hong
@sunhahong.bsky.social
2.1K followers 430 following 22 posts
Wrote Technologies of Speculation (NYUP, 2020). Writing Predictions Without Futures. right now: Stanford Humanities Centre Fellow, soon: UNC Chapel Hill.
Posts Replies Media Videos
Reposted by Sun-ha Hong
ketanjoshi.co
This Yougov survey contains an important finding: people who rely heavily on ChatGPT rate it as being "good" at tasks that we ABSOLUTELY know for sure it is very bad at

imo, this is a strong piece of evidence these heavy users simply do not ever check outputs

yougov.co.uk/technology/a...
While many critics of AI might hope that familiarity with the tools – with their frequent hallucinations and often outdated knowledge bases – would breed contempt, the opposite appears to be the case. Those Britons who use AI chatbots more frequently are the most likely to think they are good at pretty much all of the tasks listed.
April 24, 2025 at 8:13 PM
Reposted by Sun-ha Hong
ketanjoshi.co
I think we can split @hankgreen.bsky.social's question out into two categories:

1) Uses with an obvious failure mode
2) Uses with a subtle failure mode

My theory is type (2) failures are becoming EXTREMELY widespread - the ones we catch are the iceberg tip, and suggest 1,000 un-caught errors
Stanford Professor Allegedly Includes Fake AI Citations in Filing on Deepfake Bill
Professor Jeff Hancock from the Stanford Social Media Lab submitted a legal argument in support of a Minnesota deepfake bill, but it reportedly includes citations made up by AI.
www.pcmag.com
April 24, 2025 at 12:42 PM