At the heart of debates over the power of social media platforms and their implications for free expression is the question of who should shape what we see online—and whether that control can shift.

Throughout the Institute’s half-day program—“Can Middleware Save Social Media?”—those questions played out across three panel discussions about online speech and the rules governing social media platforms.

As Jameel Jaffer, executive director of the Knight Institute, said at the outset: The question is not just whether social media can be saved, but what we are saving it from—and who gets to do the saving.

Defining Middleware

Panelists described middleware not as a single product, but as a layer of tools that sits between users and platforms, shaping how information is filtered and delivered. Early examples such as ad blockers showed how users could modify the content that platforms deliver. Newer systems go further, allowing users to curate their own feeds or apply alternative algorithms.

However, that shift comes with constraints, including legal barriers that limit interoperability, privacy risks, and the reality that platforms still hold enormous structural power and control the terms of engagement. Middleware, in this sense, operates within and around existing systems rather than replacing them.

In conversation with the Institute’s Ramya Krishnan, Daphne Keller of Stanford University and Richard Reisman of the Foundation for American Innovation explored these dynamics, emphasizing how middleware can redistribute control to users without inviting direct government intervention. At the same time, they underscored that middleware’s effectiveness depends on the surrounding legal and technical environment—whether platforms are required, or even willing, to allow interoperability. Without that, the promise of user control remains contingent, not guaranteed.

Middleware’s Promise and Its Limits

The keynote conversation, moderated by the Institute’s Policy Director Nadine Farid Johnson, brought together Olivier Sylvain, senior policy research fellow at the Knight Institute and professor of law at Fordham University, and Ethan Zuckerman of the University of Massachusetts Amherst to examine what middleware can actually do.

For Zuckerman, middleware offers a tangible but limited path forward. “The problem with middleware is that it is a partial, incomplete, imperfect solution,” he said. Still, he described it as “a way of showing that another world is possible,” one that could help build the case for broader reform.

At one point, the conversation turned to whether users can realistically take control of their own online experience. Zuckerman mentioned tools that prompt reflection—like weekly screen time reports on smartphones—as evidence that users might engage more intentionally with their media environments if given the opportunity.

Sylvain challenged that premise, shifting the focus from individual behavior to system design.

“There is an underlying pathology, a structural problem that middleware does not answer,” Sylvain said, arguing that the current system is defined by information asymmetries—what platforms know, what users do not, and how that gap is leveraged to shape how content is delivered and consumed. He warned that focusing too squarely “on the user as a solution … doubles down on the problem we have now.”

The exchange crystallized a central tension: whether meaningful reform begins with empowering users to reshape their own experience, or with restructuring the systems that shape it for them.

Considerations for Policymakers

In conversation with the Institute’s Ryan Morgan, panelists Renée DiResta of the Stanford Internet Observatory, Anna Lenhart of Common Sense Media, and Luke Hogg of the Foundation for American Innovation examined the policy landscape surrounding middleware and platform governance, with a focus on what interventions could meaningfully shift power in the digital public sphere.

They discussed a range of legislative proposals at the federal and state levels that could facilitate middleware development and adoption, alongside broader efforts to strengthen data privacy protections and address the structural dynamics that shape how information is distributed and consumed online.

The discussion made it clear that middleware alone cannot resolve the deeper forces shaping online discourse.

What middleware can do, however, is expose the limits of the current system and sharpen the case for structural reform. The future of online discourse will not be decided solely by platforms or policymakers, but by whether new systems can redistribute control in ways that are both meaningful and durable.

The larger question is whether the digital public sphere can be restructured so that the power over what we see is not concentrated by a handful of platforms, but more broadly shared. Middleware offers one entry point.