There’s a lot that I can say about Emily M. Bender, but I think that a philosophy professor of mine said it best when he described her as the “cutting edge of technology and AI and linguistics and ethics.” Obviously some of her cutting-edge-ness concomitantly stems from the cutting-edge-ness of large language models, deep fakes, and 'artificial intelligence' inventions. But out of all the computational linguists, Emily M. Bender stands out to me because she's made the problem of unregulated AI pertinent and understandable to everyone—linguists, computer scientists, climate activists, lawyers, everyone. Her message about LLMs and other AI inventions is clear: we have to do something, and soon, preferably yesterday. Because there is great incentive for AI to remain unregulated at the cost of our democracy, our right to privacy and ownership over our data, our planet, and (as she calls it) our "information ecosystem."
This episode answers all the questions you've had about 'AI' technology: how is the language of an LLM intrinsically different from the language of a human? What are the legal implications of un-watermarked synthetic media? What's going on with deep fakes? How can linguists use their knowledge to effect change? And throughout it all, you'll hear Emily's wisdom and empathy radiating through her wealth of knowledge.
On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜
Baldwin: Understanding the link between joint attention and language
Patricia Kuhl TedTalk: The Linguistic Genius of Babies
