@giletcapital : We can't have trustworthy AI when human bias seeps into AI chatbots through training data. Preprocessing filters are removing crucial info, amplifying bias. As is the case here, it leads to misinformation + unreliable AI responses. Clear propaganda • TwiDoom

Martin

@giletcapital

+ Follow

Forbes 100 under 100. Reformed derivatives trader. LTCM. I occasionally tweet about rugby, macro, and memes about working from home.

ID: 705330328177659904

calendar_today03-03-2016 09:53:49

1,1K Tweet

875 Takipçi

1,1K Takip Edilen

Martin

@giletcapital

2 months ago

We can't have trustworthy AI when human bias seeps into AI chatbots through training data. Preprocessing filters are removing crucial info, amplifying bias. As is the case here, it leads to misinformation + unreliable AI responses. Clear propaganda

thumb_up_off_alt0

chat_bubble_outline0

repeat0

shareShare