Could AI help us, not replace us?

The TED Radio Hour episode explores whether AI can augment human capabilities rather than displace them, drawing on insights from Siri co-founder Tom Gruber, Century Tech founder Priya Lakhani, and Robinhood CEO Vlad Tenev.

Humanistic AI as Augmentation

Tom Gruber, who helped develop Siri after selling his startup to Apple in 2011 following a meeting with Steve Jobs, defines humanistic AI as systems designed to collaborate with and empower people, rather than compete by automating tasks. Introduced in his 2017 TED Talk, this contrasts with machine-centric AI focused on surpassing human smarts for efficiency gains. Gruber notes two industry paths: one automating white-collar work (with billions raised), the other tackling humanity's challenges to make people smarter. Early on, he built AI for non-speakers with conditions like cerebral palsy, using tiny language models to predict words from single muscle inputs, outputting synthesized voice. Siri addressed similar "handicaps," like hands-free use while driving. Now, with large language models (LLMs) like ChatGPT, AI offers conversational partners that are nearly free and vastly capable, but Gruber warns of risks like superhuman persuasion leading to manipulation. Guardrails require embedding theories of human harm and benefit into AI objective functions—what the AI optimizes for—prioritizing human flourishing over profit.