The illusion of safety: How Meta and other platforms are dismantling Trust & Safety online

When Mark Zuckerberg announced earlier this month about the changes to Meta platforms in their trust and safety policies, it beckoned responses from a lot of directions, but many focused only on the fact-checking aspects of the broader actions he was announcing.

Those of us who work in health information and digital information environment, know that it is the design of the platforms that propagates narratives that can harm health and exploit consumers through fraud, scams and profiteering.

For a nuanced discussion of the changes underway, I recommend you listen to the recent podcast hosted by Kara Swisher. She interviews three of the original architects of trust and safety on social media— Del Harvey (former Twitter Head of Trust & Safety), Dave Willner (former Facebook Head of Content Policy), and Nicole Wong (former VP and Deputy General Counsel at Google)—and they discuss the evolution and unraveling of content moderation.

Their conversation explains that the goal of trust and safety teams has always been to design safer information environments from the start—not just police the chaos after it happens. (Here’s my discussion of the situation from two years ago.)

From guardrails to free-for-all

The conversation in the podcast explains that Meta’s recent decision to curtail fact-checking and reduce ranking interventions for misinformation is the latest step in a long retreat from design and actions in support of product safety. While some interpret this as a return to “free speech,” the reality is that these decisions deliberately create a more dangerous, more chaotic, and more manipulatable online space.

The experts on the podcast pointed out that Meta’s move was not just about reducing fact-checking—it involved:

• Turning off ranking interventions for misinformation, allowing harmful content to spread unchecked.
• Scaling back hate speech enforcement, particularly against LGBTQ+ individuals, immigrants, and women.
• Reframing moderation as “censorship” in language meant to appease political interests.

These shifts reflect a larger trend in which tech companies are deliberately loosening their grip on content moderation not because it is ineffective, but because it is inconvenient—inconvenient to profit, inconvenient to power, and inconvenient to those who want to exploit these platforms to manipulate public discourse.

The false choice of “free speech vs. moderation”

One of the most important takeaways from the discussion was that moderation is not censorship—it is a necessary function of maintaining any public space. The panelists compared platforms like LinkedIn and Pinterest (which have clear boundaries about the kind of discourse they foster) with X (formerly Twitter) and Meta’s platforms, which once sought to be everything to everyone and now struggle under the weight of their own lack of purpose.

Platforms that fail to define their own boundaries become battlegrounds for exploitation—whether by bad actors spreading disinformation, organized harassment campaigns like Gamergate (all the way back in 2014), or, in the worst cases, incitement to real-world violence, such as the genocide in Myanmar fueled by Facebook’s unchecked algorithmic amplification of hate speech.

The bigger picture: Who benefits from the chaos?

The conversation touched on how figures like Elon Musk and Donald Trump have aligned themselves with the push to dismantle content moderation—not out of principle, but out of a strategic desire to control the information environment. By turning platforms into free-for-all spaces where the loudest and most aggressive voices dominate, they ensure that those who are most organized in their disinformation efforts gain the upper hand.

But even as these platforms back away from trust and safety, the speakers noted that the demand for it will not disappear. As Dave Willner put it:

“People are always going to get into conflicts online, and they will demand some resolution. The question is just how stupid the journey has to be before we get back to that realization.”

Why this matters for health

More people in public health now recognize that social media and the digital information ecosystem are harming health, but the conversation remains superficial, often focused on content moderation, fact-checking, and individual interventions like social inoculation. Improving media, science, digital, and health literacy education or age-gating use of social media or apps are approaches that are more comprehensive but also need to be complemented.

Public health cannot stay reactive in this matter because bad content and misinformation aren’t the only issue. People don’t just avoid care or seek online health advice over talking to providers because of poor moderation or content online. We also fail to ensure that credible, authentic health information is available when and where people need it.

While some argue that public health should get better at using social media’s engagement-driven tools to counter misinformation and communicate health and science, in my mind public health becoming a better influencer isn’t enough for us to ensure better use of health information and healthy decisions. We must push for deeper change—not just in platform design and regulation, but also in how health information is integrated into the digital spaces where real discussions about health, wellness, and well-being happen. If public health stays focused only on fighting bad content or on getting louder and better at leveraging ad-based and influencer-supported engagement, we lose the bigger battle—reshaping the digital ecosystem so that reliable health information is not just available but trusted and accessible when it matters most.

What comes next?

In the podcast, the guests discuss that there’s a growing push for regulation—particularly in Europe and Australia—but in the U.S., much of the conversation around online safety has been hijacked by bad-faith actors framing moderation as political bias. Meanwhile, as public trust in platforms erodes and the platforms are splintering, people retreat into smaller, private spaces like group chats or decentralized platforms like BlueSky.

But as the guests pointed out, architecture matters. The way platforms are designed determines the kind of speech they foster. The real battle is not about whether we should “censor” content—it is about whether we are willing to build digital spaces that do not incentivize harm in the first place.

As Meta, X, and other platforms pull back from moderation, the question is not just whether their platforms will become more toxic—it’s whether the rest of us will critically examine how design choices shape our information environment and push for better ones.

Because in the end, letting harmful content run wild online isn’t a defense of free speech. It’s a design failure. And that failure is a choice.