The landslide US election gives me hope for online democracy

WOOT! I’ve been staying up into the wee hours since Tuesday to mainline the surprise, glee, and renewed sense of hope sweeping the Internet after American voters pulled off something close to full-on regime change. A huge Democratic majority in congress is our best chance in years to get the federal government to help make online platforms subject to the laws of the communities they serve, not the will of their owners. Here’s my recap of tech-related results of note, and three bills that I am going to be watching in the new congress.

The biggest loser on Tuesday was, of course, the Republican Party, but the tech industry is a close second. A few fun data points:

If you don’t like big tech, it’s time for a celebratory feast. I won’t suggest a menu, because you have so many proteins to choose from.

Public outcry is no guarantee of legislative success — Casey Newton just published a must-read analysis of why the push for social media regulation failed. I am more optimistic this time around because the backlash against tech changed electoral outcomes, which is what actually gets politicians’ attention. Outgoing Texas governor Greg Abbott didn’t back off on data centers in the runup to the election out of some grand principle – he saw where the wind was blowing and wanted to protect his chances to win. If this new congress doesn’t act on big tech, they’re not going to get re-elected.

A lot of bills were proposed back during the social media hearings, and more have been suggested recently that focus on AI. The ones I like best focus on process and transparency, instead of trying to define and mandate a single safety goal in the middle of a fast-changing tech landscape. Here are the three I am most interested in seeing discussed once the new congress is seated.

This bill parallels the EU’s Digital Services Act by requiring social media companies to share more about their internal workings. It was first proposed by Stanford Law Professor Nate Persily in a Washington Post op-ed in late 2021. It was introduced to the Senate by Chris Coons (D-DE), who has reintroduced it with minor changes in each Congress since then. It has had a rotating cast of bipartisan co-sponsors, usually including Bill Cassidy (R-LA). This summary of the 2023 version covers the key elements:

I wrote a lot about transparency last year when Platformocracy was young (tl;dr: I am for it). Transparency is the foundation of platform democracy, because you can’t consent to a system you can’t understand. Expect me to write a lot more about transparency in the months to come, in hopes of seeing some real change.

Ugh. I might actually agree with Ted Cruz (R-TX) on something. Cruz introduced the TERMS act, with only Republican co-sponsors, in 2024 as a bill to stop big tech from cancelling conservatives. It would require platforms to clearly publish their acceptable-use policies, give people notice before deactivating their accounts, and publish reports about policy violations and actions against users.

While I don’t agree with Cruz’s premise that there’s an insidious anti-conservative bias lurking in Trust & Safety teams, the bill’s text would have been right at home in my series on practical due process. Here’s a snippet:

Given the source, I am not confident that there aren’t any weird, partisan booby traps in the text of the TERMS act. So I would love to see a good-faith effort at a bipartisan version of this bill in the new congress. This could be an opportunity to get a filibuster-proof (maybe even veto-proof) majority to advance authentic due process for everyone, regardless of their political leanings.

Silicon Valley congressperson Ro Khanna, who I mentioned above, announced his plans for this bill in late September. It was created after discussions with leading AI safety organizations like Model Evaluation and Threat Research, the Machine Intelligence Research Institute, and Palisade Research, so represents some of the best thinking on the issue. The full text has not been published, but it was shared with CNBC.

I am excited to read Khanna’s bill when it’s published, because it goes further than PATA or TERMS by proposing a new federal agency with an annual budget and oversight authority. PATA leans on the National Science Foundation and TERMS looks to the FTC. I understand that this was probably done to make the bills easier to pass. The problem is that putting more load on existing agencies, who don’t necessarily have the expertise in these areas or the funding needed to expand their remits, means they are going to be behind from day one, and may never catch up.

Khanna’s bill recognizes that because AI platforms will be evolving quickly for the foreseeable future, effective oversight needs a well-funded regulator who can keep pace. The AI Act’s regulator would, per CNBC, have ongoing responsibility for “creating safety regulations, a licensing system to approve model training and deployment, frontier AI model audits and setting security standards for continuous testing and effective human control.”

Ideally, a new tech oversight agency would go beyond AI to address the overarching issues of transparency and due process for all of the Internet platforms we rely on. That would probably never happen in a Congress with razor-thin margins. But Tuesday changed the math, and in January we might see some real action. For the first time in years, I am hopeful.

Ideas? Feedback? Criticism? I want to hear it, because I am sure that I am going to get a lot of things wrong along the way. I will share what I learn with the community as we go. Reach out any time at [email protected].