Vint Cerf: We're only in the "early days" of artificial intelligence
The Internet and technology pioneer spent the day at Elon on Sept. 30 and was awarded the inaugural Imagining the Internet Areté Medallion.
As the robot in the video stumbled and recovered walking through the snowy woods, technology pioneer Vint Cerf couldn't help but marvel at its capabilities, while also find the humor in its uneven gait.
"Must be the 90-proof oil it was consuming," Cerf said before a packed McCrary Theatre at Elon Friday afternoon. "You notice the kind of affinity you form because it's largely a humaniform thing."
The robot created by Boston Dynamics is evidence of the miraculous advances in technology during recent years, with software and computing chips now able to better mimic how the human mind processes massive amounts of data nearly instantaneously. Artificial intelligence has a bright, optimistic future, said Cerf, who as one of the "fathers of the Internet" and the chief Internet evangelist at Google knows how a new technology can progress and become an integral part of everyday life.
"This is early days for all of these technologies," Cerf said after sharing videos of self-driving cars and untippable, adaptable robots. "We won't see the end of this — presumably our great-grandchildren will. In the long run, I think they will do us a lot of good."
Cerf's Sept. 30 visit to Elon offered insights into the evolution of Internet technologies and a glimpse of what lies ahead. He talked about efforts to increase Internet access globally during an afternoon discussion in Whitley Auditorium and walked through advances in machine learning during his lecture at McCrary later in the day. His visit offered the opportunity to award him with the inaugural Elon School of Communications Imagining the Internet Areté Medallion, which recognizes highly distinguished humanist innovators who have dedicated their lives to positively impacting the global future.
Cerf, 73, worked with Bob Kahn in the 1970s to create the revolutionary TCP/IP protocol suite that led to the operational launch of the Internet in 1983, followed by an explosion of innovation that led to systems for email, file-sharing, the World Wide Web, WiFi, mobile networks and millions of online and mobile applications. For the past 11 years Cerf has been a vice president at Google and its chief Internet evangelist as he works to promote better global access to the Internet and innovation in connectivity.
"The most important thing is working for the global good," said Janna Anderson, professor of communications and director of the Imagining the Internet Center, in introducing Cerf as recipient of the Areté Medallion. "It's not possible to list every good deed he's done."
In his lecture, Cerf focused specifically on the evolution of artificial intelligence and the potential it presents, minimizing concerns that it could go too far. Focusing largely on the work of researchers under the Google umbrella, Cerf explained how online search has evolved to contextualize the massive amounts of content and data so that a responses to search queries are more targeted and useful.
Google's Knowledge Graph has taken that concept further by gathering data and inferring conclusions from it, Cerf said. But despite those advances, a significant portion of the Internet lies within the "Deep Web," which has no metadata that search engines rely upon to glean information, he said.
"This is kind of like dark matter in the universe — we know it exists, but we can't see it," Cerf said. "We have a long way to go to make our artificial intelligence mechanisms capable of seeing all of what's in the World Wide Web."
Advances in creating processors, software and computers that can digest information the way the human mind can have produced some milestones, most notable in the ability for supercomputers to win complex games like chess and Go played against humans. Google, IBM and others are developing processors that create neural nets that simulate the way the human mind processes stimuli and data, but there are still so many unknowns in how the process works, Cerf said.
"It's a little unnerving to think we're building machines we don't understand," Cerf said. "On the other hand, it's fair to say that the Internet is at a scale now that we don't understand it, either."
The future lies in the development of software rather than hardware, Cerf said, given that "there is no limit, there is no boundary on software. Whatever you may be able to imagine, you may in fact be able to program."
The main limitation and potential danger is in the fallibility of software programming, with the introduction of bugs and too few safeguards against failure, he said.
"We need to remember that they (robots) are made out of software, and we don't know how to write the perfect software, and we don't know how to write bug-free software," Cerf said. "The more we rely on them, the more surprised we may be when they don't work as expected."