This and all episodes at: https://aiandyou.net/ .

AI runs not just broadly across human interest, but deeply. And at every level it seems to create paradoxes. How can it be useful without power; how can it be safe with power. We want it to take over jobs, but still leave us with meaning and purpose. We deny it the possibility of becoming sentient, but how can we trust something that lacks compassion? My guest today has written a whole book on these paradoxes. Calling from Sweden is Virginia Dignum, professor of responsible artificial intelligence at Umeå University, and the author of the 2019 book Responsible Artificial Intelligence. She’s an internationally recognized expert in AI ethics and policy who has led initiatives for the European Commission, the United Nations, the World Economic Forum, UNESCO, and UNICEF, among others. And we are talking about her new book, The AI Paradox: How to Make Sense of a Complex Future.

We talk about the issues of ethical choices and power that AI raises, where the difference between humans and LLMs matters, Viriginia’s path to the work she’s doing and her book, and our trend to techno-solutionism and what it means for us.

All this plus our usual look at today's AI headlines!

Transcript and URLs referenced at HumanCusp Blog.

This and all episodes at: https://aiandyou.net/ .

Have you ever thought about your thoughts? About what or how you’re thinking? It gets real meta real fast, doesn’t it? That’s called metacognition, and humans and certain other creatures do it. But what about AI? We’re coming back to the interview with Ricky Sethi, Professor of Computer Science at Fitchburg State University, and researcher into artificial metacognition, or whether and how machines can think about thinking. Ricky’s research spans fact-checking misinformation, virtual communities, and artificial metacognition, where he focuses on designing GenAI systems that can monitor, evaluate, and regulate their own reasoning. He is Director of Research for the Madsci Network, and an Adjunct Professor at Worcester Polytechnic Institute. He has over 50 scholarly publications, and his work has been covered in outlets such as the Chicago Tribune, The Conversation, and Communications of the ACM.

We conclude the interview by talking about his research into disinformation, measuring the emotions associated with it, how clusters of models in different roles could assess AIs for lying and implement AI safety, and how AI is impacting the job opportunities in research.

All this plus our usual look at today's AI headlines!

Transcript and URLs referenced at HumanCusp Blog.

This and all episodes at: https://aiandyou.net/ .

Have you ever thought about your thoughts? About what or how you’re thinking? It gets real meta real fast, doesn’t it? That’s called metacognition, and humans and certain other creatures do it. But what about AI? Can it think about thinking? Here to help us understand this whole thing is artificial metacognition researcher Ricky Sethi. He is Professor of Computer Science at Fitchburg State University, Director of Research for the Madsci Network, and an Adjunct Professor at Worcester Polytechnic Institute. His research spans fact-checking misinformation, virtual communities, and artificial metacognition, where he focuses on designing GenAI systems that can monitor, evaluate, and regulate their own reasoning. Is that cool or what? Ricky has a bachelors in neurobiology and physics, an M.S. in physics/information systems, and a PhD in AI from UC Riverside. He has over 50 scholarly publications, and his work has been covered in outlets such as the Chicago Tribune, The Conversation, and Communications of the ACM. Recently, he has introduced the Metacognitive State Vector framework for quantifying key cognitive signals in ensembles of large language models.

We talk about how this spans computer science, neuroscience, and psychology; System One and System Two thinking come up again, with a beautiful explanation. We also talk about testing and measuring metacognition in humans and AIs – and what about dolphins?

All this plus our usual look at today's AI headlines!

Transcript and URLs referenced at HumanCusp Blog.