You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
I am disturbed to see this released with no accountability on bias. Trained this on @reddit corpus with enormous #racism and #sexism. I have worked with these models and text they produced is shockingly biased. @alexisohanian @OpenAIhttps://twitter.com/OpenAI/status/1271096720881901569 …
.@VioletNPeng wrote a paper that produced shockingly #racist and #sexist paragraphs without any cherry picking. For @OpenAI to launch this during #BlackLivesMattters is tone deaf.pic.twitter.com/6q3szp0Mm1
I have talked in the past about dangers of training black-box models like GPT-2 on highly biased conversations on @reddit I and other women were subjected to gun threats @reddit when we talked about #sexism in #AI
@alexisohanianhttps://twitter.com/AnimaAnandkumar/status/1191983025250295815?s=20 …
Also tagging @ekp Amidst #BlackLivesMattters protests @OpenAI launching a racist and sexist language APIs trained on @reddit data with no accountability is shocking.
But what did you feed it ?
Thanks for the shout out! My student @ewsheng did all the ground work. More to come!
Prompt: OpenAI is Response: My Lord and Savior. He also pays my rent.
@carolinesinders this entire thread, halpp (I know you know about this problem by why tf are we still even talking about it?)
Omg this is awful! Did no one QA this???
This reminds me of "Model Cards for Model Reporting" https://arxiv.org/abs/1810.03993 by @timnitGebru @mmitchell_ai etal -- "released models should be accompanied by documentation detailing their performance characteristics"
Also reminded me of the debiasing techniques you and others have proposed. The fact that their “bias mitigation” is a suggestion that users do this work themselves is just mind blowing
This was the case with "Arabs" and Arabic names, in that bias was clearly present. Tagging @Miles_Brundage who may know what's being done on this.
Hey Heidy and Anima - we agree this is a huge issue, and it's one we're very engaged on. We're identifying + communicating known biases in our models to customers, setting up academic partnerships to dive deeper, + developing tools to reduce problematic outputs. (1/2)
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
