The Future of Free Speech

Can AI help people change their minds? At #FreeSpeechSummit2025, Alice Siu shared research showing that AI-assisted deliberation can shift opinions just as effectively as human-moderated discussions.

People are comfortable telling AI their deepest, darkest secrets… as long as they know they’re talking to AI. At #FreeSpeechSummit2025, Alice Siu explores this interesting phenomenon and what it means for how we communicate in the AI era.

“You have no idea what the person next to you read in their news feed this morning.” At the #FreeSpeechSummit2025, Alice Siu argues that different feeds do more than shape our opinions— they prevent us from finding common ground.

DeepSeek exported censorship. Open source developers reversed it in 48 hours. At the #FreeSpeechSummit2025, @brooksben.bsky.social explained how developers modified DeepSeek’s open-source model to address censorship of responses about Taiwan, Hong Kong, and Xi Jinping.

California tried banning deepfakes in elections. Months later, politicians were making their own AI-generated attack content. At #FreeSpeechSummit2025, @brooksben.bsky.social argued that what counts as “dangerous AI content” often depends on politics, context, and who holds power.

“Trust us bro.” At #FreeSpeechSummit2025, @brooksben.bsky.social warned that the public AI companies and governments are asking the public to trust systems they cannot meaningfully inspect or scrutinize.

"It's insane." The EU government has launched a new app that lets people verify their identities to use social media. In a @reason.com interview with @nickgillespie.bsky.social, @jkosseff.bsky.social asks: Would you really want to upload your ID to the government in order to post online?

“If [the West] did not have press freedom, we would have to invent it for them." That's what a KGB operative said about its mission to discredit democracies by instigating hate against Jews and creating support for hate speech bans. @mchangama.bsky.social explains:

As countries race to regulate artificial intelligence, the language written into these laws could shape how rights, accountability, and free expression evolve for everyone.

Are deepfake bans protecting democracy, or restricting political speech too broadly? At #FreeSpeechSummit2025, Kyung Sin Park argued that Korea’s AI laws ban all electoral deepfakes, including satire, parody, and even truthful content.

“DeepSeek doesn’t just block politically sensitive questions. It generates answers that echo the party state.” At #FreeSpeechSummit2025, Ge Chen warned that AI can be used not only for censorship but for large-scale propaganda and political influence.

“AI is never treated as a neutral or technical tool.” Ge Chen explained how China’s AI regulations go far beyond content moderation, embedding censorship directly into AI systems while also using AI proactively for state propaganda.

“Digital sovereignty” proponents argue that, to protect democracy, nations must control their online spheres. But would Viktor Orbán have lost in Hungary if the Internet were under government control? @mchangama.bsky.social at @aodemocracies.bsky.social's 2026 Copenhagen Democracy Summit:

The latest assassination attempt on Donald Trump reflects a moment of heightened political tension and violence in America. But at @france24.com, @mchangama.bsky.social explains why calls to restrict speech won’t heal this division — and would actually make matters much worse.

In 1961, the French government swept the killing of dozens of Algerians under the rug. On The Gist podcast, @mchangama.bsky.social talks to @milkepescs.bsky.social about how citizens now have the technology to push back against official government narratives — but only if those tools remain free.

In 2021, Congress introduced a bill that would have exempted “health misinformation” from Section 230 protection — defined by the HHS Secretary. But would those who supported that law at the time still be okay with handing that power to an unelected bureaucrat today?

Chatbots' “hallucinating” is the least of our problems. While critics often cite AI inaccuracies as a fatal flaw, at our #FreeSpeechSummit2025, Tyler Cowen argues this is the most overrated problem in the current discourse.

We’ve already seen the dangers of governments using AI without guardrails. But at the #FreeSpeechSummit2025, Tyler Cowen argues that if we don’t focus on building a free AI ecosystem now, governments will start to control the systems we rely on.

Do you trust the current administration as the arbiter of truth? What about the next one? At a recent @cato.org event, @mchangama.bsky.social explains how bottom-up social media fact-checking systems can empower citizens to question official narratives.

AI has radically transformed how students learn. But at our #FreeSpeechSummit2025, Tyler Cowen warns: our education system hasn’t caught up yet.

If AI can beat one of the world’s top economists at his own game, what does that mean for the rest of us? At the #FreeSpeechSummit2025, Tyler Cowen explains how we can grow with AI.

At our #FreeSpeechSummit2024, @szabolcspanyi.bsky.social discussed Hungary’s crackdown on press freedom:

We built these platforms, but now they’re shaping us. It’s time to flip the script. At the #FreeSpeechSummit2025, @noupside.bsky.social says platforms must place control in users’ hands.

How can we give social media users power over their feed? @noupside.bsky.social discusses ways to reshape the relationship between platforms, users, and their phones at our #FreeSpeechSummit2025.

Free speech will always be under threat. It is the people's greatest tool and authority's greatest fear. In an interview with @gadsaad.com.web.brid.gy, @jkosseff.bsky.social discusses his new book and why we can never stop defending this freedom, no matter how well--intentioned restrictions seem.

Private platforms can set their own rules under the First Amendment. But at our #FreeSpeechSummit2025, @noupside.bsky.social explains what's in it for the government when it tries to sway platforms' moderation decisions.

🚩 Your feed isn't "neutral." At our #FreeSpeechSummit2025, @noupside.bsky.social explains that every time you open an app that ranks content, someone else is deciding what is important for you to see.