AI Content Chat (Beta) logo

Journalism This new button could trigger an audit from a third-party site equipped with sophisticated AI tools to assess the article on the fly. Are the statistics the article cites verifiable? Appropriately contextualized? Who are the sources it quotes, and what addi- tional information about them would be helpful? How does the story fit into its topic’s larger context? What are the sources of any images, video, and audio the article incorporates? Are these elements genuine or synthetic? Providing this level of scrutiny for every article a news orga- nization publishes might seem like overkill, and maybe it is. Certainly a human-powered system like this would be too time-consuming, too costly, not feasible. AI gives us new superpowers, however, and we should apply them ambitiously. In a world where hard-to-detect disinfor- mation could mix freely with carefully reported, methodically verified information, we must do everything we can to make the good stuff easy to identify. What kind of impact might this have? For starters, it could quickly help sort the world’s information into two basic categories: that which proactively encourages evaluation and verification, and that which doesn’t. Of course, the problem with fake news is not just fake news, it’s that so many people want fake news because it supports what they already believe to be true. Even so, what’s the downside to creating a more visible culture of informational transparency and accountability, where the “ingredients” of at least some news articles become as legible and easy to check as the nutri- tion label on a can of soup? 93

Impromptu by Reid Hoffman with GPT-4 - Page 100 Impromptu by Reid Hoffman with GPT-4 Page 99 Page 101