Last week, I attended TrustCon, a global conference in San Francisco for trust and safety professionals. While I was there, I liveblogged two sessions and I wrote some thoughts on the field-forming process trust and safety seems to be going through:

Liveblog: Trust & Safety x Responsible AI: Bridging the Gap

Liveblog: Keynote Speaker Catherine Teitelbaum

For whom is trust and safety professionalizing?

In addition to these posts, I wanted to also share some of what I learned from other panels I attended but haven’t already written about. 

  1. Regulation of social media has already come, but we still don’t know where it is heading. 

For example, the EU’s Digital Service Act requires some platforms to conduct risk assessments and audits of their moderation systems and to publish regular transparency reports, starting this year. Platforms are in the process of complying with these requirements for the first time, and regulators have not generally given extensive guidance on what exactly they are looking for. Platforms are trying to pass a test without a study guide, and regulators are waiting to see what they produce before committing to sharing any basic expectations. I imagine we’re likely going to see an iterative process of developing best practices over the next few years in which platforms look at each other’s work and the Europe Commission gives guidance in the form of fines to signal the need for correction next year.

Outside of the EU, many countries are passing their own online safety laws. They are aware that the difficulty of complying with dozens of laws with very different requirements could “splinter” the internet — a concern that is not new. That’s why we’re seeing efforts like the Global Online Safety Regulators Network emerge to facilitate coordination across jurisdictions. This network is relatively new, so we still don’t know how it will influence new laws and regulations among its seven member states, but it could come to play a significant role in setting standards for regulation internationally.

  1. Trust and safety professionals share a desire for more coordination across platforms. 

A huge frustration of trust and safety professionals is that they might ban a bad actor from their platform, but that person may continue causing harm on other platforms undetected. There are very few avenues for sharing information among trust and safety teams across companies. The Global Internet Forum for Countering Terrorism allows for hash sharing of specific images related to terrorist content or CSAM. And last year, an organization called the Tech Coalition announced a program called Lantern for sharing signals regarding child safety investigations across platforms. But there is an appetite for more of this. 

This includes a lot of discussion around the need for more open-source tooling in trust and safety. While different from signal sharing, access to a common repository of tools and content classifiers that are not tied to a specific service provider could contribute to sharing best practices and procedures across teams. 

These conversations reminded me of legal scholar Evelyn Douek’s argument about content cartels — that too much shared information and standards for removal across the industry could harm freedom of expression if those systems do not include necessary transparency and accountability mechanisms. It’s an argument anyone working on signal sharing should be familiar with.  

  1. Speaking of tooling, the trust and safety vendor space continues to grow.

One reason I attended TrustCon was to interview founders of trust and safety vendors for my dissertation. I had no shortage of companies to talk to, and I learned a lot about the various services and tools they provide. Most of them are not trying to capture the entire trust and safety workflow but are instead aiming to slot in a specific piece of the trust and safety ecosystem. 
This includes a big focus on tools that don’t make moderation decisions themselves, but offer a platform for managing data and the workflow of making decisions. It makes sense that (especially) smaller and medium sized companies should not have to each build their own bespoke moderation management system internally. But, as I’ve written about before regarding the tool many volunteer Reddit moderators use, moderation tools are not merely about scale. They structure work and decision making in ways that shape the kinds of governance that are possible. These tools aren’t neutral, but most conversations around transparency overlook them and instead focus on content decisions. As the vendors who build them grow in prominence, we should recognize the influence they will begin to exert over moderation practices.