🛫 I thoroughly enjoyed my quick 2-day trip to Singapore. I couldn’t resist a bit of mischief by titling my keynote at Singapore’s AIMX Conference, “Navigating Frontier AI Safety: The Science of Responsible AI in Australia” 🇦🇺, given it was just a day before the UK AI Safety Summit. 😜

Setting cheekiness aside, our scientific core message emphasizes:

• 🛠 System-level engineering practices, guardrails, and supply chain accountability

• 🎯 RAI for frontier model-based systems (rather than just feeding more data to the model and hoping for the best)

• 🗣 Empowering decision-makers & minority voices

Despite the raging debates about x-risk distracting from the more immediate concerns, the fundamental scientific questions are sometimes the same. We need to understand the systems better, harness and control them more effectively, and improve our risk measurement across all dimensions of responsible AI.

Slides (dropbox): Navigating Frontier AI Safety: The Science of Responsible AI in Australia

𝐑𝐞𝐬𝐞𝐚𝐫𝐜𝐡 at CSIRO’s Data61: https://lnkd.in/gyzjE4-i
𝐁𝐨𝐨𝐤: https://lnkd.in/gsQz5swy