Online platforms' effects on public health, safety and democracy

This discusses a 2026-05-06 interview with Yael Eisenstat about the impact of online platforms on public health, safety and democracy. A video and 29:00 mm:ss podcast excerpted from the interview will be added when available. The podcast will be released 2026-05-16 to the fortnightly "Media & Democracy" show syndicated for the Pacifica Radio Network of over 200 community radio stations.

It is posted here to invite others to contribute other perspectives, subject to the Wikimedia rules of writing from a neutral point of view while citing credible sources and treating others with respect.

2026-05-06 interview with Yaël Eisenstat about how some online platforms have been degrading public health, safety and democracy

29:00 mm:ss excerpts from 2026-05-06 interview with Yaël Eisenstat about how some online platforms have been degrading public health, safety and democracy.

Yaël Eisenstat discusses the impact of online platforms on public health, safety and democracy. She is currently the Director of Policy and Impact at Cybersecurity for Democracy (C4D), working on policy solutions for how to hold social media and other online platforms accountable for their effects on public safety and democracy. Previously, she was Vice President at the Anti-Defamation League (ADL) Center for Technology & Society (CTS). She was a Facebook election integrity head in 2018 and later became a whistleblower, speaking publicly about the dangers to democracy stemming from the company's decisions and products. She has held other other positions protecting democracy including as an intelligence officer, diplomat, and White House advisor. C4D contributed to the recent March 24, 2026, jury verdict in a civil case against Internet companies in New Mexico.

Eisenstat is interviewed by Spencer Graves.

Eisenstat's work

Eisenstat's work includes a TED Talk and an SXSW panel in 2020 and a 2024 research report on tech platforms and political violence.

2020 TED talk

In Eisenstat's (2020) TED talk, she said that around 2015 she began to notice that she was losing the ability to engage with others who were thought differently. Conversations with others in the US were becoming more difficult than conversations she had had as a CIA officer and diplomat drinking tea and talking with outspoken anti-Western clerics and suspected terrorists in Africa. Many of those engagements began with mutual suspicion but none degenerated into shouting or insults. In some cases she built collaboration on areas of mutual interest. Her most powerful tools were to listen, learn and build empathy. Most of her contacts wanted to feel heard, validated and respected. But social media companies like Facebook incentivize inflammatory content contributing to a culture of political polarization and mistrust. This generates revenue for Facebook and similar companies that make money from clicks, "because the shortest path to a click is anger or hate", in the words of Facebook whistleblower Frances Haugen, interviewed 2024-08-19 for this Media & Democracy series. When Facebook asked Eisenstat in 2018 to lead their work to support global elections integrity for political ads, she agreed. She left six months later, speaking openly about Facebook's inability to meet its responsibility to secure elections, subsequently documented, e.g., in the thousands of internal Facebook documents that Haugen released to the Securities and Exchange Commission and The Wall Street Journal in 2021.

2020 SXSW panel

Eisenstat was part of a "panel about the Future of Tech Responsibility" for the 2020 South by Southwest festival. The festival was cancelled due to COVID-19, but the panel was held virtually. This panel included a discussion of Section 230 of the Communications Act of 1934, as amended by the Communications Decency Act of 1996. It was "written before platforms such as Facebook, YouTube and Twitter existed" -- written while Google was a research project by Stanford PhD students Larry Page and Sergey Brin. Section 230 includes, "No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider." Eisenstat says that it's time to revisit Section 230, to demand accountability where Internet companies promote or suppress information based on the content while protecting web freedom otherwise. This is similar to the recommendations of Dean Baker that when Internet companies make money by promoting information differentially based on content, they should be liable as are legacy media under the US Supreme Court decision in NYT v. Sullivan (1964). In other cases, they should be treated as common carriers like telephone companies.

2024: Tech platforms and political violence

More recently, Eisenstat et al. (2024a, b) are insisting that, "Tech Platforms Must Do More to Avoid Contributing to Potential Political Violence". The New York Times had reported that, "a steady undercurrent of violence and physical risk has become a new normal," particularly targeting public officials and democratic institutions. A survey from the Brennan Center found that 38% of election officials have experienced violent threats. They attributed these threats primarily to tech platforms and gave seven recommendations in four themes "congruent with any number of papers that academics and civil society leaders have published over the years." They said that platforms

• must develop robust standards for threat assessment and engage in scenario planning, crisis training, and engagement with external stakeholders, with as much transparency as possible.
• should enforce clear and actionable content moderation policies that address election integrity.
• should enforce their rules uniformly, not exempting politicians and other political influencers.
• must clearly explain important content moderation decisions, ensuring transparency especially when it comes to high profile accounts.

They hope that increasing demands for accountability will prompt platforms to act more responsibly and prioritize the risk of political violence both in the United States and abroad.

Highlights

These excerpts are rushed, lightly edited for readability, and may not be in final form. The ultimate authority on what was said is the accompanying video.

Most important issues

When asked about the most important things she wanted to communicate to this audience, Eisenstat replied,

“

There's been a debate for a number of years now on whether or not or how social media companies and online platforms are affecting our democracy, whether or not they're contributing to the rise in extremism, to threats to our elections, to democratic backsliding. These have been really important questions we've been interrogating for a number of years now. I think we've really arrived at a tipping point now, especially with some of the cases that we have seen happening in this space recently, that really show that some of the things that many of us have been saying for years are really starting to be proven, both by the actions and statements of some of the leaders of these companies themselves, or some of the documents that are showing up in court cases now.

But I like to start by debunking some of the myths that surround these conversations. No, I am not saying that social media is the end all, be all cause of all of our grievances, and that democracy would be a perfect thriving thing if not for social media. I have never said that. That is not actually the point of the conversation.

The question is: What role have these companies been playing, whether it's undermining elections, perhaps bringing more extremist fringe views to the mainstream and contributing to radicalization of certain people and populations and also as tools for some authoritarian regimes and leaders to truly influence not just the space in which we make sense of our information and in which we engage, both in political debate and democracy and all of it, but fundamentally, what effect are they also having on our safety, our public health, all of these issues?

And so I think, you know, we're at a real tipping point where what was long viewed as some activists screaming from a mountaintop, some academics proving things through their scientific rigor, some journalists exposing things. I think we're actually at the point now where enough of the public has understood that these issues really matter, and so let's figure out how to course correct and hopefully fix some of them.

”

Recommendations

Graves then asked for her recommended course corrections. She replied,

“

We have allowed a very small handful of companies and tech leaders to have an unimaginable amount of control over how information flows, how political campaigning is communicated, how we are actually making sense of the information that underpins a healthy democracy. We have allowed that to be controlled by such a small handful of people. The irony is the original promise of the Internet was to create a much more open system, where the traditional small set of gatekeepers would be disrupted. ...

I love the idea of democratizing both access to information and access to be the voices of issues. All of these things, all of these aspirations, were wonderful ideas.

But let's be very clear, that is not how it is shaped up. Once a very small handful of people figured out how to both dominate this industry, scale their ideas recklessly, ensure that they could do so with all the legal protections, more than any other industry has ever been granted in place and with no accountability for the outcomes of their actions.

There's no one silver bullet. ... The public needs to understand how these technologies are affecting them. Because even let's say you wanted legal and legislative changes, which I do, ... you need a constituency that believes that that change has to happen to pressure their lawmakers to actually undertake that process of fixing some of these issues.

There's public education of really helping people understand, and I think that we've made some really great strides in that. If you just look at these recent court cases, more and more of the public -- they don't need to understand how an algorithm works. ... They just need to understand there is something that is not working here. I've seen enough to know I don't like how it's addicting my child or I don't like how it's making conversations so extreme that nobody feels safe to disagree with anyone anymore. I don't like how it's allowing foreign state actors to intervene in our elections in a way that was never so easy and, frankly, cheap in the past.

They only need to understand that they don't like that and then say that something has to change. They don't have to be the ones to figure out exactly what that change is. And then it's on lawmakers. It's on courts, but also public pressure to the companies. ...

”

Migrating from for-profit to non-profit social media and Internet companies?

Graves then asked if she had suggestions about how to migrate from for-profit to non-profit social media and Internet companies? She replied,

“

It is really challenging right now. I want to be clear, I still use these companies. I would be a hypocrite if my takeaway was to tell everybody to get off of Instagram, TikTok, Facebook, whatever platform people are using. ...

I don't want to imply that I think everybody needs to get off all of these platforms. At the end of the day, it's very hard to put the toothpaste back in the tube. These companies are part of every ounce of our being. I'll give you an example.

A man I know around 57 years old never joined Instagram. Never used it. And then his child went off to college, and the parents group around all of the freshmen that were all in a group together created a group on Instagram to communicate with each other about their kids' experience and whatnot. And that person at 57 years old had to join Instagram if he wanted to be part of the parents group for his kid that just went off to school.

I hate that. I wish they would have figured out a different way to create a parents group. But at the end of the day, telling people "You just need to get off these platforms" is a bit of an unfair fight, when these platforms are so ubiquitous and such a part of our everyday life.

That said, people can understand how to use it in a more healthy way. ... I don't see a nonprofit social media space growing to the scale of something like Meta to have the network effects of basically engaging almost everybody on the planet. ...

If you look at Bluesky as the answer to X, that's not really the answer to what you're talking about. That's still a very niche community. ...

I would just like to get to the point where there's enough understanding that the status quo is not okay, and that we get to figure out how to differentiate product safety from this sort of false narrative that every single thing about a social media company is just about free speech, and therefore it's untouchable. Hopefully, some of these recent cases have started to prove that there are real business decisions, deceptive practices, design features that are not necessarily about content, but that cause real harm in the world.

These are the things we need to start grappling with, as opposed to telling everybody how to just get off the platforms.

I would love to give you a recipe for how to rely more on nonprofit social media. I hope that happens someday, because at the end of the day, as long as these companies entire existence is to make money, and the only way they can truly believe they're going to make money is through this surveillance advertising business model. These issues are going to be very hard to solve, but I don't think you're going to find the nonprofit platforms catching up yet.

”

Frances Haugen

Graves noted that he had interviewed Facebook whistleblower Francis Haugen in 2024, who said,

“

The shortest path to a click is anger or hate.

”

Graves asked for Eisenstat's thoughts on that. She replied that Haugen was referring to Meta's business model:

“

When a social media company's business model is predicated on being able to sell you personalized, targeted ads, they want to keep you on the platform as long as humanly possible, so they can hoover up as much of your data as possible so they can sell this idea to an advertiser. ... For example, they know exactly which sneaker you're going to click on, as opposed to just sending that ad to like everybody on the platform. That doesn't sound so harmful. Cool: Maybe I'll get an ad for something that's interesting to me.

The point about the anger and the emotion, which is something actually I really dug into in my TED Talk in 2020, ... that algorithm is coded with a goal in mind. Let's be very clear, it is a human being who is making a decision when designing that algorithm. It is not some mysterious thing that created itself. ...

This is not something unique to social media. I spent the first half of my career in the field, on the ground, before social media existed. I was one of our government's top counter extremism officials. I was working with communities susceptible to extremist messaging in the early 2000s to help figure out how to inoculate them from being exploited by extremists who were going to prey on those vulnerabilities to then radicalize them in the future.

And what we're seeing online is the same exact thing but at a much huger scale. The algorithms know what is going to keep you engaged. And we all know that's going to be emotional content. We all know that's usually going to be something that gets your blood racing. So it's usually going to be something gets you really upset, really angry, really outraged. And that's what's going to keep you going and engaging and sharing and commenting. They're not doing that because they care about that content. They're doing that because the more you share, click, engage, the more they're able to sell you and your data and your behavioral profile to an advertiser. ...

That is fundamentally the problem with this business model. It's why all the externalities are what's really concerning to me.

I don't know that Mark Zuckerberg purposely said ever. -- I doubt he did -- "I want to make sure that more QAnon content is recommended to people." Or, "I want to make sure that more and more people are watching these misogynistic content creators talking about how to do terrible things to women." ...

There is enough evidence that that is exactly what his algorithms have been incentivizing, and because it has been incentivizing that content, there is a real world consequence of taking what was once more fringe and making it more and more mainstream. And what the most important thing to understand is what happens online has real consequences offline.

”

Externalities

Graves noted that Eisenstat mentioned externalities and asked if she had suggestions about internalizing those externalities? She replied,

“

"Some of them fall into areas that unfortunately are harder to solve because they're not actually illegal, but they are causing real harm. There are different categories. I look a lot at how these platforms are helping to fuel things like political violence, extremism, radicalization -- things that are, I think many would agree, quite harmful to our democracy and our society. Again, repeating what I said earlier, I'm not saying social media is the end all, be all singular reason that these things are happening.

But as others have said (I am not alone in this), they are certainly the gasoline being poured on the fire. ...

Some of the documents that Francis exposed are really important in proving some of these points. ... There was an article written called "Carol's Journey to QAnon". ... It was based on Facebook's own internal research at the time. It showed that Facebook internal researchers had set up this account, named her Carol, made her a mom from, I think, like the Midwest somewhere, had her start to follow a few things online, including, I think, a few politicians.

But it didn't start with, "I am following QAnon" and all these crazy like right wing conspiracies. And it showed that within a very short period of time, the recommendation engines were recommending her QAnon content.

We've replicated studies like that outside of Facebook. We did a study called something like the amplification of antisemitism and hate. It also showed that the recommendation engines themselves were recommending children, teenagers who weren't even searching for conspiracy theories or antisemitic or other kinds of hate filled content, were recommending these things to them.

But the reason I bring up the documents that Francis exposed is that proves that the company knew it. ... They knew these things were happening. That's a very critical point ... going back to the idea of a business model that does end up incentivizing the most extreme and salacious content. ... I think that "Carol's Journey to AQnon" is the perfect example.

We started looking also at like the profiles of some of the people who commit mass violence in the United States. ... It has been skewing more and more towards young men or committing acts of political violence. I'm not going to sit here and say every single one that's ever happened. But if you look at the profile online of a lot of these people ... are spending a lot of time online in these conspiracy theory sort of rabbit holes ... .

After the January 6, 2021, insurrection on the Capitol, you started to look at the online activity of certain individuals that were involved and implicated. You can see it. You can see that they were going more and more into very extreme content, reposting some very harmful conspiracy theories. ... Whether it's X, whether it's TikTok, whether it's Instagram, I get that they are not personally responsible for the White supremacist or Neo-Nazi out there who wants to spread hate.

But I cannot accept that they bear no responsibility for how their own engines that they have designed not only amplify that, but actually connect people to it. Actually recommend it to children. ...

Actually recommend a 57 year old predator online. "Oh, these are your likes. Here's a 13 year old girl you might want to connect with."

And, yes, that has been proven to happen. Both the investigation that the New Mexico Attorney General did and some other independent civil society and academics have done research in this space showing that the recommendation engines connect predators with children.

These are the things that, to me, are inexcusable, that we allow companies to do with absolutely no repercussions.

”

Graves interrupted, "No: There are big repercussions. They make money doing it." Eisenstat agreed,

“

Correct, and those are the things that we need to change. ...

A lot of the work that I'm involved in right now, or have been for years, is really trying to disentangle this idea of what about what online really is speech, and I believe in some of the core spirit of this original law from 1996 called Section 230, which, at its core, preemptively immunizes these platforms for the speech of their users ... while also immunizing them so that they can engage in content moderation. ...

That second part people forget about a lot. ...

What we've been trying to do for years ... are disentangling what is truly product design that's not actually about somebody else's speech, but is about trying to elicit a particular behavior from what is third party speech so that everything doesn't get thrown out before even getting to see the light of day in court.

And so that is why these two cases, Los Angeles and New Mexico, these two big cases against Meta and Los Angeles was also against Google.

These were the first two cases that actually went to a jury, because for so many years, they've been preemptively immunized by Section 230 and what was different in these cases, and you can read some of our work on this.

We filed an amicus brief for the big MDL, the multi district litigation. We've written a lot about these cases. Some of our research underpinned some of it ... showing that these are products ... using computer science and technological research to prove that some of the products are designed 100% to have nothing to do with content itself, but to elicit a behavior.

You can just read their patents, which is what we did. If you read our amicus brief, we did a lot of patent analysis. If you read their patents behind things like infinite scroll, or behind things like their intermittent reward, variable rewards, they patent this technology. They're very clear of what goal they are trying to achieve with it. Regardless of the content, those design features, those technologies, if they are, in the end, contributing to an illegal activity or to harm, that's what we should be addressing.

I might hate some of the racist, homophobic, antisemitic speech that's happening online. That's not the core of what I'm trying to go after. ... Even if I hate it, it's free speech, it's First Amendment protected. It's part of what makes us an open society,

But how a company itself is designing its tools to keep us on the platform, to elicit a certain behavior, to allow people to even monetize the worst of the worst. And then what harm is that creating? That's the important question.

”

Can Zuckerberg be prosecuted for perjury?

Graves noted that Zuckerberg testified in one of these trials and asked if he could be prosecuted for perjury? Eisenstat replied,

“

That's an interesting question. ... I'm not a lawyer, so I don't know the exact answer to that. I will say that for years, we have heard, especially senior executives from Meta lie about what they know about their own products. We've heard it in congressional testimonies. ... I wasn't at the trial in Los Angeles. I don't know whether he lied or not on the stand ... .

Part of why it's so important for a case like this to go to trial is because ... then you get to go into discovery. And what happens during discovery? You get a bunch of documents from internal communications. And the internal communications are showing exactly what they knew and what they decided to do with what they knew.

And my understanding from these cases, and part of the reason the juries returned the verdicts they did, was it was proven very clearly where they keep lying about what they knew, where they keep lying about the fact that they knew their products are harming children or other issues involved there, and whether or not they decided to do anything about it.

Whether he can be held for perjury, I don't know. ... I don't know if he actually lied on the stand in Los Angeles. I have heard multiple Facebook senior executives say untrue things, both in the media and in congressional hearings, but I'm not sure about in the case itself.

”

In sum

Graves then noted that we are about out of time and asked for final words for the audience. Eisenstat replied,

“

For parents right now, it can feel very overwhelming, because you keep getting told that social media is the most terrible thing in the world, and predators are going to stalk your children, and it's going to create terrorists around every corner. I'm not saying that's the case.

I do think it can be quite dangerous, and especially for parents. They really need to get to understand the platform. Have an open conversation with their children about how they're using it. Explain to them the healthy ways to use it. Engage as a parent with your child's experience online.

That said a lot of adults aren't using it in a healthy way either. ... Not everything that happens on social media is terrible. It's not. I use it still.

Government's job, whether it's regulators, whether it's courts, whether it's legislators, their job is to protect the vulnerable from illegal activity, from exploitation, while hopefully also allowing innovation to thrive. ...

The amount of money and power that is at stake here is unprecedented in US history. A lot of people don't understand why it has taken so long. I've been screaming about this since I left the company in 2018 and I decided to go public as a whistleblower and speak about what I saw happening there. People were speaking out before I did, and people continue to take really serious risks to try to expose what is happening in these companies.

But the amount of money and power at stake is why it is so hard to find the solutions. There are so many constituents yelling at legislators, "If you do this, you're going to break the whole internet. If you do that, like the world will cease to exist."

But I would just ask people to keep supporting those who are truly doing the work. It is a long battle, but I think we are starting to see real success. The ultimate goal is not to shut down every social media company. The ultimate goal is to figure out what a safer online experience looks like and what accountability looks like when something unsafe happens.

”

The need for media reform to improve democracy

This article is part of category:Media reform to improve democracy. A summary of episodes to 2025-11-15 is available in Media & Democracy lessons for the future.

Discussion

[Interested readers are invited to comment here, subject to the Wikimedia rules of writing from a neutral point of view citing credible sources and treating others with respect.]

Bibliography

• Bobby Allyn (25 March 2026). "Jury finds Meta and Google negligent in social media harms trial". NPR. Wikidata Q139572103. https://www.npr.org/2026/03/25/nx-s1-5746125/meta-youtube-social-media-trial-verdict. 
• Musadiq Bidar; Laurie Segall; Dan Patterson; Michael Kaplan; Jessica Kegu (25 October 2021). "Facebook internal documents show execs knew platform spread misinformation and failed to act at times". CBS News. Wikidata Q139719521. https://www.cbsnews.com/news/facebook-documents-misinformation-failed-to-act/. 
• Center for Technology & Society (16 August 2023), From Bad To Worse: Amplification and Auto-Generation of Hate, Center for Technology & Society, Wikidata Q139719742
• Yael Eisenstat (August 2020), Dear Facebook, this is how you're breaking democracy, TED, Wikidata Q138844363
• Yael Eisenstat (2021). "Section 230 Revisited: Web Freedom vs Accountability". Cornell Tech. 13 May 2020. Wikidata Q139568755. https://dli.tech.cornell.edu/post/section-230-revisited-web-freedom-vs-accountability. 
• Yael Eisenstat; Justin Hendrix; Daniel Kreiss (2024a). "Preventing Tech-Fueled Political Violence: What online platforms can do to ensure that they do not contribute to election-related violence". The Bulletin of Technology & Public Life. 22 May 2024. Wikidata Q139571027. https://assets.pubpub.org/zn130hqk/Preventing_Tech-Fueled_Political_Violence-51716391441318.pdf. 
• Yael Eisenstat; Justin Hendrix; Daniel Kreiss (2024b). "Tech Platforms Must Do More to Avoid Contributing to Potential Political Violence". Tech Policy Press. 22 May 2024. Wikidata Q139571163. https://www.techpolicy.press/tech-platforms-must-do-more-to-avoid-contributing-to-potential-political-violence/. 
• Yael Eisenstat; Christopher Murray (8 July 2025), C4D partners with legal and social media experts on amicus brief in MDL case against Meta, Cybersecurity for Democracy, Wikidata Q139720052
• Morgan Lee (31 January 2026). "Undercover investigation of Meta heads to trial in New Mexico in first stand-alone case by state". Associated Press. Wikidata Q139719943. https://apnews.com/article/meta-facebook-trial-new-mexico-social-trial-facebook-instagram-whatsapp-d8b812efd001e5cabbef9e1a47143226. 
• Katie McQue (24 March 2026). "Meta ordered to pay $375m after being found liable in child exploitation case". The Guardian. Wikidata Q139572337. ISSN 0261-3077. https://www.theguardian.com/technology/2026/mar/24/meta-new-mexico-jury. 
• Blake E. Reid (4 September 2020), Section 230 of… what?, Wikidata Q139570229