Back in 2018, Ben Shneiderman sat down with Alyssa and made a sharp prediction: that AI would quickly become the go-to “catch-all” for fixing the internet’s biggest problems—especially things like algorithmic bias and the way content is given to users. He pointed to Mark Zuckerberg’s congressional testimony, where Zuckerberg confidently claimed that AI would eventually solve the issue of people being exposed to polarizing or uncomfortable content. Fast forward to today, and it’s clear that he was onto something. AI has become the tech industry’s favorite buzzword—every company is racing to slap "AI-powered" on their products, often with more enthusiasm than results. It’s been positioned as the magic fix for misinformation, bias, hate speech, and just about everything else that goes wrong online.
But while hype has soared, reality hasn’t really caught up. AI hasn’t eliminated algorithmic bias; in some cases, it’s even made things worse by reinforcing existing patterns in the data it learns from. Content moderation (in my opinion) is still a mess, and recommendation engines continue to create echo chambers that amplify division rather than reduce it. The promise of AI as a neutral, corrective force has proven far more complex than many tech leaders anticipated in 2018. Shneiderman was right to be skeptical—AI has been treated like a silver bullet, but the world's still figuring out how to aim it.