Google Public Policy’s Post

AI is too important not to regulate, and too important not to regulate well. To keep up momentum, we're sharing some principles for smart regulation.

As teams across Google work to make AI even more helpful for our users, we're setting a new standard for responsible innovation with a full end-to-end approach. Our most recent paper looks at our current process for pre-launch and post-launch controls — the AI Responsibility Lifecycle: 1) Research to understand the capabilities and limitations of new technologies 2) Design and test continuously, using safety thresholds to fine-tune models 3) Govern by assessing risk, focusing on mitigations and testing for safety and fairness 4) Share best practices so organizations of all sizes can benefit from your learnings The paper offers insight into how we developed each step in this process, with recent examples and practical tips for how others might apply these learnings. As we continue learning more through our lifecycle approach to AI responsibility, we commit to sharing knowledge with the whole ecosystem. To learn more about our approach, view the paper at https://goo.gle/3RJmOou

Proud to see Quill.org featured on the TODAY Show for the work they’re doing to increase millions of students’ reading and writing skills through their free AI-powered tool. One of the most-used AI literacy tools in the US, Quill has already benefited 9M students — 5M+ of whom are at Title I schools. Looking to the school year ahead, Quill is aiming to double their impact as a member of Google.org’s Generative AI Accelerator program. See more: https://goo.gle/45pEOJX

We’re helping build a strong #cybersecurity workforce and combat #cyberattacks by supporting 15 new cybersecurity clinics at higher ed institutions across the U.S. Learn more about the Google Cybersecurity Clinics Fund and new Cyber Clinics here: https://goo.gle/3VbAjOL