Conversation

Adam Strandberg
@strandbergbio
Good luck making a word blender for the entire internet not say anything bad
Quote Tweet
OpenAI
@OpenAI
·
Generative models can display both overt and diffuse harmful outputs, such as racist, sexist, or otherwise pernicious language. This is an industry-wide issue, making it easy for individual organizations to abdicate or defer responsibility. OpenAI will not. 3/8
Show this thread
5
13
153
niftierideology
@niftierideology
AI safety researchers then [swole doge]: How do we keep AI from destroying the human race? AI safety research now [pathetic doge]: my language model keeps saying pee pee poo poo
2
4
46