Earlier this month, INFER hosted a special forecasting event about the future of AI, attended by select INFER forecasters and nearly two dozen forecasting enthusiasts who work at Google, joining in a personal capacity.
This event was part of INFER’s Mission: AI Advancement – a forecasting challenge to examine the current AI-boom, the future of generative AI, and its impact on society. A handful of INFER forecasters who completed the challenge were invited to attend the private event that included a group forecasting session.
Two of the guests presented their perspectives on the current AI boom: Seth Blumberg, a behavioral economist, and Pietro Kreitlon Carolino, a software engineer. They focused on generative AI broadly from economics, machine learning, and engineering perspectives, including how forecasting can be useful as AI technology develops.
INFER Pro Forecaster Spencer Henderson, shared some reflections after hearing from the guest speakers and having subsequent discussion with fellow forecasters: “With AI, we often lack rigorous definitions about what even constitutes intelligence, or what artificial general intelligence would even look like. Good forecasting can fill in this gap in understanding. I think we should expect more of the same in the future: impressive AI developments that change the world, but that humanity is more than capable of coping with.”
Following speaker remarks, INFER forecasters and Googlers were divided into break-out groups to collaborate and forecast on the question: How many autonomous vehicle collisions will the California DMV record for October, November, and December 2023 combined?
This question was chosen due to recent policy changes allowing the operation of self-driving taxis (or “robotaxis”) across San Francisco. Advances in artificial Intelligence, robotics, and sensory technology have sparked a revolution in self-driving vehicles. However, concerns about safety and distrust of AI may hamper progress.
Using INFER’s “request for response” capability, we asked each break-out group to submit their forecasts. This allowed us to review an aggregate forecast from each group in real-time as they were submitted, and directly compare forecasts and perspectives between groups.
“The meetup was a great chance to observe how other people make decisions in their forecasting process and how they factor in uncertainties or unknown variables. The session was information dense, and a lot of value was exchanged during our discussions,” shared INFER forecaster Heramb Podar.
We appreciate all who joined and participated in the challenge to think collectively about the future of AI. Be sure to watch out for new AI-related questions and events coming soon to INFER!