Why Zuckerberg's AI vision misses the mark with a skeptical public

Mark Zuckerberg recently published a lengthy manifesto about personal AI, focusing on the "personal superintelligence" systems Meta AI is developing. The essay, which expands on themes he's shared before in the Wall Street Journal and on earnings calls, paints a hopeful picture of AI's societal benefits. But it also repeatedly highlights potential downsides, and critics argue it does little to ease public concerns about the technology.

The timing is notable. Social media platforms and their leaders face significant distrust in the U.S., with a recent survey showing that 64 percent of Americans believe social media has harmed democracy, and a similar share think it should face stricter regulation. A court also recently fined Meta $567 million over child safety issues. Given this backdrop, many people are already wary of tech executives' promises about AI's positive impact.

Zuckerberg's essay includes broad claims about how giving everyone access to superintelligence will lead to positive outcomes through natural checks and balances. He also describes a future where everyone has a personalized tutor with a PhD in every subject, and where superintelligent lawyers make the legal system fairer. However, critics point out that AI tutors are already being used mainly to help students avoid doing homework, and that legal AI could just add more complexity or encourage frivolous lawsuits.

He also discusses Meta's freemium model, suggesting dynamic auction pricing for compute access. While the idea of broad access is reasonable, surge pricing for essential AI tools could be a poor user experience. The essay, critics say, fails to acknowledge real harms and does little to build trust, especially when compared to other tech leaders who more openly discuss risks and safety measures.

Source: TechCrunch