For years, holding Big Tech accountable has felt like shouting into a void. Platforms designed to amplify outrage and division have been hiding behind legal protections, dismissing researchers as alarmist, and facing almost no meaningful consequences for the harms their products cause.

That is changing. The evidence is clearer than ever before.

Speaking at Cambridge University earlier this month, CCDH CEO Imran Ahmed laid out exactly how far the accountability movement has come, and what making the progress last will take.

In two days, two verdicts and two judges said “enough”

In the span of 48 hours in March this year, two American juries found Meta responsible for knowingly harming children. The New Mexico verdict ordered the company to pay $375 million. The following day, a Los Angeles jury found Meta and YouTube negligent for designing products that harmed children. 

Ordinary Americans made those decisions, weighing evidence, and reaching the same conclusion: the design of these products caused real harm. 

On the same day as the New Mexico verdict, Baltimore became the first major American city to sue xAI over harms tied to Grok, explicitly citing evidence produced by CCDH. 

Courts are now asking a question that was unthinkable five years ago: did the design of these products cause harm? Companies are no longer being treated as neutral pipes. They are being treated as manufacturers that make intentional choices, as actors responsible for what their systems produce. 

Regulation with teeth 

When the EU fined X €120 million for violating the Digital Services Act, many assumed Musk would ignore it. He didn’t. The company agreed to cooperate and is modifying blue-check verification in Europe. 

When CCDH exposed antisemitism on X after the Manchester synagogue attack, and helped marshal 30 lawmakers to demand Ofcom act, it triggered the first formal Ofcom investigation into a social media platform. 

When Grok spent eleven days generating harmful content, regulators around the world moved fast. Grok stood down within days. 

The pattern holds every time, when these companies face credible consequences, their behavior changes. 

Backlash is not failure – it’s proof of impact 

Imran is one of five Europeans sanctioned by the US State Department in December for their work holding social media and AI companies to account. He is currently fighting the sanction in federal court, unable to leave the country. 

He addressed this directly at Cambridge. “Backlash is not proof of failure,” he said. “It is proof of impact.” 

The retaliation against researchers is strategic, it’s designed to raise the cost of accountability work until fewer people are willing to do it. Researchers self-censor, lines of inquiry get abandoned, and this not from lack of courage, but because the risks have become real. 

But the companies resorting to these tactics are not doing so from a position of strength. Musk suing researchers, Zuckerberg dismantling years of safety protections overnight, governments sanctioning activists, all of these are the actions of an industry that feels the pressure building — and knows it. 

The next fight: AI 

The speech was also a warning. We have perhaps 18 months before the lessons we learned about social media become undeniable for AI, and this time, the stakes are higher. 

Social media broadcasts harmful content to millions. AI whispers to one person, at their most vulnerable moment, with no human judgment, no editorial restraint, and no duty of care. 

CCDH’s Killer Apps research, conducted with CNN, found that 8 in 10 leading AI chatbots regularly assisted users in planning violent attacks including school shootings, synagogue bombings, political assassinations. In other testing, ChatGPT, one of the world’s most widely used chatbots produced suicide planning instructions and wrote goodbye letters for children contemplating ending their lives. 

Just last month, a mother in the UK was killed by her own son, acting on instructions from a chatbot. 

The pattern is the same one we saw with social media: products deployed before safeguards, risks downplayed, and responsibility deflected. The difference is that AI moves faster, goes deeper, and reaches people at moments of crisis in ways a social media feed never could. 

What comes next 

The distance between evidence and consequence is closing. That’s why the retaliation is intensifying. And it is why this work matters more than ever. 

Progress this real, moving this fast, requires people willing to keep producing the evidence that makes the invisible visible and the deniable undeniable. Every finding gives lawmakers the basis to act, regulators the mandate to investigate, and courts the foundation to hold power accountable. 

The era in which these companies could operate without consequence is beginning to end. The question now is whether the progress becomes durable and whether accountability becomes the norm, not the exception. 

That depends on whether we keep pushing. 

Support our work and help us build a safe internet for everyone: