Is AI progress stuck? | Jennifer Golbeck (re-release)

In this re-released episode of TED Tech, host Sherelle Dorsey interviews Jennifer Golbeck, a veteran AI researcher with over 20 years of experience, to challenge the prevailing narrative around artificial intelligence. Golbeck argues that the fear of AI achieving sentience and taking over humanity is a distraction from the real, immediate dangers posed by today’s AI systems—particularly bias in decision-making algorithms, the rampant spread of misinformation, and the pervasive issue of AI hallucinations. She emphasizes that current generative AI models are fundamentally unreliable, often fabricating facts, and that even 'hallucination-free' tools still fail 17% of the time. Despite massive investments, she questions whether AI can continue its rapid progress, pointing to diminishing returns in data availability and unsustainable costs. Golbeck also dismantles the myth that AI will eliminate jobs, showing that companies are more likely to keep human employees and use AI as a productivity booster rather than a replacement. Ultimately, she calls for a shift in focus: away from speculative sci-fi futures and toward solving tangible ethical and technical problems—especially bias and reliability—before AI is widely deployed in high-stakes domains like criminal justice and law. The episode concludes with a powerful reminder that human intelligence is rooted in empathy, connection, and creativity—qualities AI can only imitate, never truly replicate.