Synthesia reposted this
Fun to appear on BBC Question Time last night! Today, a disproportionate amount of AI debate is on the "what-if's" and philosophical/existential questions around what could happen with AI. Ultimately, no one knows – neither me, OpenAI, Hinton, Googlers, researchers or anyone else. We'll have to find out. The only thing we know is that historically we've been incredibly bad at predicting the future. It's human nature to focus on one technology in isolation and imagine all the ways it could go wrong, not right. Spending so much time and attention on unknowns is causing us to neglect the problems that are present and solvable right now, which we can and should be tackling. We'll learn more by doing than dreaming, and we do have a choice in how we shape the future. Philosophizing is also important and we should obviously consider long-term scenarios – but everything in proportion. Social media algorithms reward attention-grabbing soundbites over practical realities, unfortunately.