Building agents is like writing code in a programming language where individual instructions (LLM calls) are stochastic. The challenge is to figure out programming abstractions and error correction mechanisms that will enable reliable code on top of such an unreliable primitive.

Offering a sort of gift: If you manage someone in Ethical/Responsible AI, please give them license for regular mental health days off. To step away from a headspace that often requires saturating in the darker side of things. I'll explain some of why. 🧵

Urgent information for any government departments currently using these tools to summarise and synthesise research for policy briefings.

In a year when we're all so concerned about AI-generated misinformation, the call may coming from inside the house...

🚨****PLEASE APPLY***** for Financial Support for ACM FAccT '24

🚀Travel funds are absolutely available.

Due April 12:
- Estimated travel + acco costs
- For students, institutional funds available to you if any
- Read form ASAP for other info
facctconference.org/2024/scholarsh…

Today, the U.S. government is establishing new rules for its own use of AI, and they’re a HUGE deal

Buckle up folks — we’re going to unpack how they work and why they’re so good 🧵

blog.mozilla.org/netpolicy/2024…

So excited to announce an event The GenLaw Center has been working on!

We're discuss the misconceptions b/w the technical capabilities of evaluating generative AI, and what policymakers and civil society want...

April 15th Georgetown Law Tech Institute, and live on zoom:  dc-workshop.genlaw.org

.Center for Democracy & Technology submitted comments regarding the risks, benefits & potential policy approaches to “open foundation #AI models” in response to @NTIA’s request:
cdt.org/insights/cdt-c…

Nominations for participation from federal employees are being sought for Princeton University's AI Policy Precepts Spring 2024 Program, which will explore the future of both predictive and generative AI. Arvind Narayanan Matthew Salganik Zach Vertin citp.princeton.edu/programs/princ…

Reading this new report from the FTC and it looks like a fantastic resource for other regulators trying to increase their AI/tech capacity:
ftc.gov/reports/ftc-st…

.Center for Democracy & Technology joins Mozilla & nearly 50 orgs + scholars in a letter to Secretary Gina Raimondo, urging her to protect openness & transparency in #AI in response to NTIA public consultation to examine risks/benefits/potential policy approaches re: open AI models:
cdt.org/insights/cdt-j…

Call for nominations: Princeton University AI Policy Precepts. Non-partisan program on ideas, opportunities & risks around AI. For policymakers in Congressional offices & Federal agencies in DC. Teachers are Arvind Narayanan, Mihir Kshirsagar & me. Nominate by 3/29 citp.princeton.edu/programs/princ…

Are you an EU citizen, and interested in AI governance?

These roles at the new AI Office are incredibly important and interesting. Not only will you be creating evals of advanced AI systems, you can help enforce sanctions and stop these systems from being deployed. Details below

Also at #SOTU2024 , President Biden called on the federal gov’t to advance responsible AI. That work is well underway, with the The White House’s #AI Executive Order laying out requirements for gov't-wide coordination to protect #CivilRights , #privacy , and democratic values in #AI use.

💥 New paper 💥

We discover a form of covert racism in LLMs that is triggered by dialect features alone, with massive harms for affected groups.

For example, GPT-4 is more likely to suggest that defendants be sentenced to death when they speak African American English.

🧵

AI company policies like account bans and legal threats can chill independent evaluation.

Today, we are releasing a paper and an open letter (signed by 100+) calling for safe harbors for independent evaluation.

Letter: sites.mit.edu/ai-safe-harbor/

Paper: sites.mit.edu/ai-safe-harbor…

The 14th Annual Privacy Papers for Policymakers will take place at the Capitol Visitor Center in Washington, DC!🏆

We’re joined by keynote Senator Peter Welch, awardees Claire McKay Bowen, Albert Fox Cahn🦊 (@[email protected]), & others, along with discussants Edgar Rivas, Miranda Bogen, & more.

eventbrite.com/e/14th-annual-…

'Less Discriminatory Algorithms' — which I wrote with Emily Black, Pauline Kim, Solon Barocas, and Ming Hsu — was selected as one of the winners of Future of Privacy Forum's Privacy Papers for Policymakers Awards.

you can read it here: papers.ssrn.com/sol3/papers.cf

📍New job alert📍
The brand new AI Collaborative (an initiative of The Omidyar Group) is looking for a Director of Operations to oversee all aspects of grantmaking, investing as well as the overall functioning of the org. More info and how to apply here: apply.workable.com/ai-collaborati…

.Center for Democracy & Technology AI Governance Lab's Miranda Bogen joins #InTechnology to discuss the intricacies, implications, and impacts of the recent US Executive Order on #AI with Divyansh Kaushik, Chloe Autio, and Camille Morhardt.
Listen: cybersecurityinside.libsyn.com/181-deep-dive-…

👀NEXT WEEK: CDT AI Governance Lab's Miranda Bogen is joining The Brookings Institution's Joshua Paul Meltzer + Dr. Chinasa T. Okolo in panel discussing last week's UK #AI SafetySummit & how it fits in with broader trends toward cooperation in the global policy landscape. RSVP:

cdt.org/event/ai-safet… #AI