Peter Mattson, founder and president of MLCommons, emphasizes the importance of AI safety by stating, "AI is a really young technology, and AI testing is a really young discipline. Improving safety benefits society; it also benefits the market."
MLCommons is launching the AILuminate benchmark, which evaluates AI language models on their potential harms, focusing on threats such as hate speech and promoting self-harm.
The new benchmark AILuminate tests large language models against over 12,000 prompts related to various harmful behaviors, providing scores from "poor" to "excellent."
Donald Trump’s potential policy changes regarding AI could shift the landscape for AI safety, as he may eliminate the measures introduced under President Biden that promote responsible AI use.
Collection
[
|
...
]