Now at #MuslimsInML workshop 🌙. @abidlabs on how GPT3 has signifant anti-muslim bias, particularly in reference to violence. 1/n 
And the bias on display about @MarzyehGhassemi. Needed to completely distract with pizza to get the model to generate text unrelated to violence and bias. 2/n
And the @AnverEmon led us to consider all the roles of AI in rosk based assessments and international finance and charity 🤩 Best quote: 'who gets to be conservative. If you are brown and muslim you're a national security threat.'🔥🍕 3/n
Latest at #MuslimsInML we are learning about The digital enclosure of Uyghur and Kazakh societies and the internal settler colonialism they are experiencing. This talk is an incredible eyeopener 🍕 4/n
And now we are talking about low resource languages and experience of a women as an arab/african/woman in AI/tech. So much inspiration and insight; by Mona Diab 🤩🖤 5/n 🍕
Last talk of #MuslimsInML is by @DiaKayyali. On tbe Syrian Archive, the sudanese archive, content moderation and how ML is implicated in different ways and the move away from human in the loop decisions. 6/🍕
Wisdom from the panel from @mo_norouzi and @abidlabs: because muslims are hesitant to put data about themselves, they get degined by data that's already there about them. 🔥
• • •
Missing some Tweet in this thread? You can try to
force a refresh
