The rise of AI and self-driving cars
Whether you buy in to the hype or fear, AI is everywhere these days.
To ground folks into why this is the case, Ira A. Fulton Schools of Engineering Professor Subbarao Kambhampati gave a tour de force overview on the last four decades of AI during a standing-room-only AAAS session on “The Rise of AI: Status, Prospects and Thresholds.”
“AI is bigger than you suppose,” Kambhampati said. “And it’s bigger than you can suppose. Everybody is falling over each other to say their consumer products have AI in it — even electronic toothbrushes. But naysayers say it will destroy the world.”
Kambhapmpati says one of the key reasons AI has become the topic de jour is to understand the contrast between the development of human intelligence and artificial intelligence. In human development, a baby first shows perceptual intelligence, recognizing the faces and voices of their parents; then emotional intelligence, managing anger and tantrums; then social intelligence of interacting with others; and finally, reasoning and cognitive intelligence.
“If you think about AI’s progress toward intelligence, AI almost went in the opposite direction,” Kambhampati said. “Back in the 1980s, it was rule-based systems. In the 1990s, reasoning, and Deep Blue vs. Kasparov in chess, and now, perceptual intelligence with speech, language and image. Perceptual abilities made AI come to us. And it leads to all sorts of off-the-wall worries that people make up in their own lives.”
That contrast explains a lot of what has been developing in AI. We could program the computers about how to play chess, and the rules of companies. But we could not program the computers themselves, we had to let them learn, and slowly recognize images and speech. And these perceptual learning aspects of AI are only as good as the algorithms programmed by people, who can be prone to biases.
“Intelligence is the ultimate dual-use technology, and it can be used for good and bad. We can generate fake videos, fake text, and in the future, this may boggle the mind,” Kambhampati said.
Just this year, China has developed the first AI English anchor who provides real fake news — but don’t buy into the concept of killer AI bots quite yet.
“AI systems are now doing pretty well on perceptual tasks, but we are still far from real intelligence. AI technology is here to stay, but there are significant positive impacts it can bring,” he said.
Kambhampati joked that above all, percepts AI needs a new type of serenity prayer to live by: “Humans, grant me the serenity to accept the things I cannot, use data to learn the things I can, and the wisdom to know the difference!”
Future of driverless cars
One game-changing societal application for AI is autonomous vehicles. Recently named as a top 10 academic think tank, the Washington, D.C.-based ASU Consortium for Science Policy and Outcomes (CSPO) — under the leadership of faculty members Mahmud Farooque and Jason Lloyd — organized a conference on the future of driverless cars, which has the potential to eliminate half of current jobs. In 2017, they helped launch the first project focused on autonomous vehicle citizen forums that will soon be expanded to 25 cities in the U.S. and Europe.
“We deal a lot with emerging technologies, and there are uncertainties on the technology side and uncertainties on the societal-impact side,” Farooque said.
The first findings of the study were reported at AAAS from University of Maryland colleague David Tomblin, who showed that citizens cared most about how autonomous vehicles would impact their families and social lives in surveys of citizens from both a big city, Baltimore, and rural Cumberland, Maryland.
“The most important thing were the social issues,” Farooque said. “And it flipped so quickly, like, ‘I’m going to get that time back that I take my parents to the doctors, my kids to the soccer practice. But at this same time, that’s my only contact with them during the day. I’m going to lose that.’ And what if they go to the casino instead of the doctor’s office? Now, you can see that the hopes and the fears are very closely linked, and it’s all about societal things, not necessarily the technology side.”
While Baltimore citizens grappled with issues like free time, the rural residents of Cumberland saw AI as a huge infrastructure opportunity to become more connected.
“They were more hopeful in the rural areas because they see that they are currently disconnected from mobility choices,” Farooque said. "There is one Uber driver in that town. If it’s automated, I can reconnect. But would it may certain problems worse? We have an opioid crisis. With all of these cars without drivers, we don’t know where they are coming and going.”
According to Farooque, the CSPO teams’ autonomous vehicle survey results could be summarized into three main issues.
“One bucket is autonomy: How much control do I want to give, and how much control do I want to retain? Number two is the economy and jobs: Am I going to lose, or am I going to gain? What’s going to happen? And number three is safety, but not just safety about saving lives, but is it going to be safe for me to get in one? What’s going to happen in my everyday life?”
The ASU CSPO-led large societal forums on autonomous vehicles will take place in Phoenix, Washington and Boston in May. Thereafter, it will expand to 22 other cities in the summer.
AAAS 2019 honorees
At the conference, Fulton School of Biological and Health Systems Engineering faculty Samira Kiani became the very first ASU recipient of the 2019–2020 Alan I. Leshner Leadership Institute Public Engagement Fellows, given to 10 researchers in this year’s theme of human augmentation. These scientists and engineers were chosen for having demonstrated leadership and excellence in their research careers and an interest in promoting meaningful dialogue between science and society.
“The topic I’m working on is very timely with engaging the public,” Kiani said. “There are so many implications with CRISPR, gene editing and how it is changing the future. And it’s not just in health, but every aspect of our future: climate, food, everything.”
During her fellowship year, Kiani wants to have the opportunity to extend the reach of her documentary, “Code of the Wild.” In collaboration with director and filmmaker Cody Sheehy (Rhumbline Media), Kiani will also work on an initiative called Tomorrow Lab, which aims to foster a creative and engaging dialogue about rapidly developing genetic technologies, how they transform the future and what it means to be human in these reimagined societies. The online learning educational platform will bring together all of the information that exists on gene editing and CRISPR into a highly appealing visual learning tool.
“I thought this was the perfect opportunity to use filmmaking and the creative arts to bring to the forefront the type of conversations that will eventually come up," she said. "As we go forward, more and more conversations are needed because of the recent Chinese gene-edited babies and to see what is actually happening. In my lab, we are developing tools that allow for the safer delivery of CRISPR for gene therapies, and I’m very excited about these technologies. CRISPR has benefits for clinical adult therapies, and beyond.”
Erika Camacho, an associate professor in the School of Mathematical and Natural Sciences at ASU's New College of Interdisciplinary Arts and Sciences, received the AAAS 2019 Mentor Award on Feb. 17 at the conference.
In addition, three ASU faculty — Bertram Jacobs, Hal Smith and Huan Li — also attended the 2019 AAAS Fellows ceremony in recognition of their career contributions to science, and were honored by other past ASU AAAS Fellows and colleagues. Jacobs and Smith were recognized for their pioneering efforts in biological sciences, while Liu was honored for his research in information, computing and communication.
Top photo: Emily Santora, undergraduate global health senior in the College of Liberal Arts and Sciences, and Barrett, The Honors College, prepares to give her poster session talk at the AAAS meeting.