The AI community has just started to grapple with the public risks associated with the publication of high-stakes research. By studying the Meltdown and Spectre case, we can learn how thoughtful norms played out after a real-life sensitive discovery. medium.com/partnership-on

Show this thread

hudson Retweeted

𝚊𝚕𝚎𝚡 𝚌𝚛𝚊𝚗𝚣

@alexhcranz

Feb 2, 2021

Some news this morning.

137

315

2,871

What can the #AI community learn from controversial avian flu research? A lot. "[The H5N1 controversy is] worth studying for anyone whose work could pose public risks." Read PAI Research Fellow

@j_asminewang

's blog post to learn more:

medium.com

What the AI Community Can Learn From Sneezing Ferrets and a Mutant Virus Debate

Lessons on publication norms for the AI community from biosecurity

AI cannot be for the benefit of people and society if concerns about AI ethics cannot be shared and debated freely. We as a field can and must do better. We support and are grateful for the deep work and leadership of

@TimnitGebru

in the AI ethics field. (1/3)

Excited to announce new research aimed at learning what’s driving the problem of high attrition on #AI teams — and how to fix it. Interested in sharing your thoughts? #diversityandinclusion ✍️ →

partnershiponai.org

Beyond the Pipeline: Addressing Attrition as a Barrier to Diversity in AI - Partnership on AI

It's difficult to specify what you want an AI to do, but it's nearly impossible to specify everything you want an AI not to do. That’s why we’ve created a benchmark for SafeLife with

@weights_biases

—to measure & improve safety in reinforcement learning:

partnershiponai.org

Introducing the SafeLife Leaderboard: A Competitive Benchmark for Safer AI - Partnership on AI

Show this thread