“I wonder when it will be that when you call a phone number for help that you will be notified, by regulatory law, that you are talking to a machine,” said Eric Horvitz Ph.D. ’91, M.D. ’94.
His is just one of the many questions that will be addressed in the new 100-year-long study on artificial intelligence (AI) that Stanford will host beginning this year.
Horvitz, the former president of the Association for the Advancement of Artificial Intelligence (AAAI) and now the Managing Director of Microsoft Research, spearheads the project — titled One Hundred Year Study on Artificial Intelligence, or AI100 for short. The study would track the rise of technology and the ways it influences society from now until 2115.
Pamela McCorduck, an author and scholar of AI, praises the project’s ambitions in the context of existing AI research.
“In more than fifty years of witnessing AI develop, to me, AI100 is as important as the first program that could ever be said to think,” McCorduck wrote in a statement.
“Cognitive computing, machine intelligence, is a powerful technology, maybe one of the most powerful the human race has ever produced, and it will become more powerful as time goes on,” she added. “If we handle it carefully, it has the potential for conferring great benefits, both at the personal level and the planetary level.”
According to Horvitz the project has its roots back in 2009 when he commissioned a study through AAAI called Presidential Panel on Long Term AI Futures.
“We had three committees about advances of AI and their influences on people and society,” Horvitz said.
“Now, a number of things happened in the past five years. There are rough edges we can be proactive about to make things go better,” he added when discussing the motivation for the new Stanford-based project.
Horvitz reached out to President John Hennessy, who he knew as a computer science colleague, and found him to be positive about the possibility of Stanford hosting the study.
The AI100 Standing Committee, which will oversee the project and will be chaired by one of its members, Barbara Grosz at Harvard, is composed of seven members including professors from universities throughout the United States and Canada. This includes Russ Altman, a Stanford professor of bioengineering, genetics, medicine and computer science, and fellow computer science professor Yoav Shoham. The committee will have rotating membership.
The project will be structured around studies conducted in five-year intervals.
The study’s framing document lists 18 topics that the study will likely consider including warfare, privacy, loss of control and communication understanding and outreach.
Horvitz explained that the current topic list is only part of an initial framing and that topics may change over time.
“The idea is not just to study these things but to be proactive, to provide guidance, to be ahead of the wave, to suggest new research areas that will be important, to make recommendations to government advisory bodies. To help guarantee better futures,” Horvitz said.
“You can see this is a very worthwhile endeavor and I’m very proud it’s going to be based at my alma mater Stanford,” he added.
One area of interest that is likely to grow is AI’s increased role in education. Horvitz cited online education efforts like Coursera as an example of an opportunity where AI can use the research results of digital classrooms for personalized education.
“For example, there’s a possibility of developing personalized models of a student’s cognition, that capture their learning style. Leveraging these models in intelligent tutoring systems may be extremely valuable,” Horvitz said.
Overall Horvitz believes the future is bright for the development of AI, especially over the long-term span of 100 years.
“I look forward to seeing our competencies in AI evolving and believe we’ll see its positive influences in many aspects of life,” Horvitz said.
Contact Zachary Brown at zbrown ‘at’ stanford.edu.
This post has been updated. In a previous version, it incorrectly stated that there were two committees in paragraph eight — there are three. Paragraph 10 mischaracterized the process of bringing the study about — Horvitz brought the idea to Hennessy, who was supportive, rather than brainstorming the idea together. Paragraph 11 misidentified the leader of the standing committee as Horvitz — it is Harvard’s Barbara Grosz. The two final quotes in the article have also been replaced to correct an editing miscommunication. The Daily regrets these errors.