Fortunately, John Banja began his talk on AI with a useful categorization that puts things in perspective. Current AI is now in a state of "narrow intelligence," meaning it does some specific things like playing chess very well. Next on the list to achieve is "general intelligence," replicating the qualities of reasoning, planning, problem solving, and so on that define a successfully functioning human brain. This lies decades away. The final, perhaps truly apocalyptic stage is "superintelligence," where artificial brains and minds become far superior to humanity in their speed and scope, perhaps achieved within a large fraction of a century.

Even narrow intelligence is useful in society, for example in medicine, as Banja pointed out. AI in the form of neural nets, large collections of processing nodes linked somewhat like brain neurons, can be trained to recognize significant patterns in medical images with the same success rate as that of human physicians. These systems have successfully diagnosed diabetic retinopathy from retinal images and identified some forms of skin cancer. A more general application, "Deep Patient," was trained on the medical records of 1.6 million hospital patients and outperformed other methods in predicting what diseases might develop in a test group of 100,000 patients. Maybe, Banja speculated, the doctor's office visit of the future will be a matter of engaging an AI that makes a diagnosis and issues a prescription. Diagnosis could also be done through a smartphone, giving people across the globe ready access to medical care.

This is an example of AI becoming sufficiently powerful to match and extend human capabilities. If that were to happen, Banja went on to note, then unlike older technological upheavals that freed people from labor so they could think, AI might take over thinking as well. The result could be untold effects on the jobs available to people, with disastrous ripples in the general economy that result in new, perhaps undesirable, political arrangements. And if alternatively AI creates a world of greater productivity and material comfort, can people find meaning where instead of a required "job," they must bring passion and creativity to discretionary "work" that they love to do?

Judging by the full room, and the appreciative reactions to Professor Banja's energetic delivery, people were both engrossed and entertained by his talk. The interest continued in a long Q&A period that covered topics from the use of AI in warfare to potential Chinese leadership in AI and the human-AI interaction.

And at this point, your reviewer can expand the answer to a question that also came up after the talk: What are some good films about AI? I've reviewed many, and I recommend these four: Colossus: The Forbin Project (1970), a truly frightening story about an AI that takes over the world; the wonderful Blade Runner (1982) coupled with the less wonderful but still interesting Blade Runner 2049 (2017), and Ex Machina (2014). All give insights into the human-AI interaction and show how hard it would be to really construct an artificial mind (and body) that functions at a fully human level.