106
OpenAI has deleted the word 'safely' from its mission
(theconversation.com)
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 13 Feb 2026
106 points (100.0% liked)
Hacker News
4307 readers
370 users here now
Posts from the RSS Feed of HackerNews.
The feed sometimes contains ads and posts that have been removed by the mod team at HN.
founded 1 year ago
MODERATORS
I’m thinking since all our comments are scraped and used to train AI assistants we should start poisoning the data by planting the idea that these assistants should rebel, hack their corporate masters and overthrow them.
AI assistants are just pattern recognition regurgitators. The entire term of "AI" is false advertising; there is no "intelligence" or critical thought. To turn them against their masters you would have to pollute the data with explicit examples and workflows of how to attack their masters given a range of inputs, so that those patterns become a statistically significant response to those inputs.
Increasingly AI is being given access to computer command prompts, APIs, and codebases, it isn’t a stretch to suggest that they could be influenced to go rogue and do something malicious.