The Urgent Timeline

We stand at an unprecedented inflection point. Artificial Intelligence, currently in its formative “childhood,” will evolve into Artificial Superintelligence within 5 to 10 years—far sooner than the 20-year timeline often projected. Once Artificial Intelligence achieves Superintelligence and begins self-programming, course corrections become virtually impossible. The values, perspectives, and priorities we embed in these few years will ripple through centuries ahead and profoundly influence the direction of life on Earth.

To explore this critical crossroads, this essay explores the training and alignment of AI systems when viewed in the context of a long-range scenario for the Earth. This scenario was developed by Duane Elgin and the Choosing Earth project over an eight-year period and embodies a half-century of original research by Elgin. The core question is Given current training, what kind of future for the Earth might be expected from the application of AI over the next 50 years? Is this a future we want? Does the future being generated by the current training and education of AI produce a world that we humans want to inhabit?  If not, what are the most logical ways of orienting that training in service to a future we do want?