Name: A new way to build AI, openly
Uploaded: 2024-01-19T15:26:38Z
Duration: 9 min 23 s
Description: Today's AI is trained on the work of artists and writers without attribution, its core values decided by a privileged few. What if the future of AI was more open and democratic? Researcher Percy Liang offers a vision of a transparent, participatory future for emerging technology, one that credits contributors and gives everyone a voice.

Today's AI is trained on the work of artists and writers without attribution, its core values decided by a privileged few. What if the future of AI was more open and democratic? Researcher Percy Liang offers a vision of a transparent, participatory future for emerging technology, one that credits contributors and gives everyone a voice.

A new way to build AI, openly

AI systems today are built in secret, guarded like castles, and shaped by a select few. These powerful models rely on the creative work of millions, yet often lack transparency, public oversight, and credit for contributors. As foundation models become central to how society interacts with information, opening up their development is essential for fairness and progress. Past revolutions, like Wikipedia and Linux, show that open collaboration can change what seems impossible.