AI, Truth, and the Challenge of Provenance
Ben Savage, founder and CEO of Veritas Automata, presented at SREday New York 2025 and talked about how the line between fact and fiction is getting blurry with AI generating everything from text to video. If a machine produces something that looks real, how do we verify its origin? Did it come from a human? Was it altered? And if AI generates misinformation, is it really “lying”?
The ability to run powerful models locally, without oversight, adds to the challenge. While AI services often include safeguards, unrestricted tools can be used in ways that are harder to track. Watermarking, source attribution, and retrieval-augmented generation (RAG) are a few ways to maintain transparency, but they aren’t perfect solutions. As AI capabilities grow, so does the responsibility to build systems that prioritize accuracy and accountability.
To learn more, watch the full presentation on the Veritas Automata channel.
Видео AI, Truth, and the Challenge of Provenance канала Veritas Automata
The ability to run powerful models locally, without oversight, adds to the challenge. While AI services often include safeguards, unrestricted tools can be used in ways that are harder to track. Watermarking, source attribution, and retrieval-augmented generation (RAG) are a few ways to maintain transparency, but they aren’t perfect solutions. As AI capabilities grow, so does the responsibility to build systems that prioritize accuracy and accountability.
To learn more, watch the full presentation on the Veritas Automata channel.
Видео AI, Truth, and the Challenge of Provenance канала Veritas Automata
Комментарии отсутствуют
Информация о видео
10 марта 2025 г. 23:18:38
00:02:46
Другие видео канала
