Op-eds
WIRED.
Meta Ran a Giant Experiment in Governance. Now It’s Turning to AI.
Meta ran a ~6500-person, 32-country, 19-language deliberative process to inform its policies, and I discuss some of the potential and problems of their approach, as both a thought partner to the team and observer of the process.
July 18, 2023
WIRED.
Red Teaming Improved GPT-4. Violet Teaming Goes Even Further
About how reducing harmful outputs isn't enough. AI companies must also invest in tools that can defend our institutions against the risks of their systems—and governance.
March 29, 2023
OneZero.
Platforms Are Still Flying Blind When It Comes to Societal Impact — Better Metrics Could Help
Shows how we can use metrics to better understand how platforms are impacting people and changing over time.
January 10, 2021
MIT Technology Review.
Making deepfake tools doesn’t have to be irresponsible. Here’s how.
Provides guidance for synthetic media tool developers, demonstrating how they can mitigate the potential for misuse. Suggests actions that academia, industry, government, media, and funders can take to incentivize this.
December 12, 2019
Bloomberg.
How to Stop Misinformation
An approach to threading the needle of privacy/encryption and protection against misinformation, focusing particularly on the context of India and WhatsApp.
January 25, 2019
Harvard Business Review.
Is Your Company Ready to Protect Its Reputation from Deep Fakes?
Why the business community should care about “deep fakes”, and why we all need to work together on mitigation.
November 8, 2018
The Washington Post.
Opinion : What’s worse than fake news? The distortion of reality itself.
High level overview of the societal impacts of advances in audio and video manipulation (“deep fakes”), and what can be done about it.
February 22, 2018
Selected Reports and Papers
Preprint: 'Optimizing for What?' Symposium.
Bridging Systems: Open Problems for Countering Destructive Divisiveness across Ranking, Recommenders, and Governance
January 24, 2023
NeurIPS Foundation Models for Decision Making Workshop.
Elicitation Inference Optimization for Multi-Principal-Agent Alignment
October 5, 2022
Belfer Center for Science and International Affairs.
Bridging-Based Ranking: How Platform Recommendation Systems Might Reduce Division and Strengthen Democracy
May 17, 2022
Harvard Belfer Center.
'Towards Platform Democracy: Policymaking Beyond Corporate CEOs and Partisan Pressure
October 18, 2021
CPX Policy Brief.
'Contextualization Engines’ can fight misinformation without censorship
April 14, 2021
AAAI/ACM Conference on AI, Ethics and Society.
The Deepfake Detection Dilemma: A Multistakeholder Exploration of Adversarial Dynamics in Synthetic Media
February 11, 2021
NeurIPS AI for Social Good Workshop.
The tension between openness and prudence in responsible AI research
December 14, 2019
The Thoughtful Technology Project.
Reducing malicious use of synthetic media research
Explores considerations relating to the malicious use of machine learning systems, and implications for potential release practices around such research.
July 27, 2019
Center for Social Media Responsibility.
Iffy Quotient: A Platform Health Metric for Misinformation
October 10, 2018
Other Writing
Newsletter (Substack).
‘Platform Democracy’—a very different way to govern big tech
November 16, 2022
Harvard Belfer Center.
Holding Platforms Accountable Is Not Enough. We Need A ‘Compass’ For Social Technologies
November 9, 2021
Medium.
How Unbundling and ‘Literacy Friction’ Might Influence the (Facebook) Oversight Board’s Decision on Trump
April 15, 2021
Medium.
When we change the efficiency of knowledge operations, we change the shape of society.
April 1, 2021
Alliance For Securing Democracy.
Deepfake Myths: Common Misconceptions About Synthetic Media
June 14, 2019
Medium.
On-Device Context: A defense against misinformation in the encrypted WhatsApp
January 30, 2019
Medium.
Even in September of 2015, hyperpartisan news sites had massive engagement on Facebook
October 20, 2017