image title
ASU roboticist discusses current technology and hit HBO show "Westworld."
ASU prof studies AI, machine learning, other topics involved in 'Westworld'
December 1, 2016

ASU roboticist Heni Ben Amor discusses artificial intelligence, machine learning, human-robot interaction

"Westworld" — the hit HBO show about a technologically advanced, Western-themed amusement park involving synthetic androids — riffs on questions about humanity, artificial intelligence and science.

The robot saloon women and gunfighters in the park exhibit intelligent behavior indistinguishable from a human.

Is anything like the tech in the show in the pipeline? Are roboticists thinking about the issues raised by creating something that looks exactly like a human? Should we make something as smart as ourselves?

To find out, we talked to Heni Ben Amor, a roboticist at Arizona State University and an assistant professor in the School of Computing, Informatics, and Decision Systems Engineering. Ben Amor’s research focuses on artificial intelligence, machine learning, human-robot interaction, robot vision and automatic motor skill acquisition.

Which is a woman and which is an android? (Answer at the end.)
Photo courtesy Heni Ben Amor.

 

Question: What are your initial thoughts about the show?

Answer: Personally I think it follows the typical trend in Hollywood to have a bleak outlook on what’s going to happen and how robotics is going to influence our society. In other countries, for example Japan, depictions of robots are more positive and have a slightly different attitude. Still, as a fan, I appreciate their artistic value.

Q: How far are we from creating what you see in the show?

A: Very far away. I think there are some underlying assumptions that are too far away from what we have at the moment. One of the things depicted there are sentient robots, robots that actually have a very good idea of what they are and what the environment is and what the emotional state of a human partner is. All of these things are not tackled by robotics at the moment. We are too far away from that, especially this aspect of having sentient robots with self-awareness and a goal or a mission that is their own. …  I think the current state of the art is the animatronics you see at Disneyworld, puppets that look fairly humanlike. They impress you the first couple of seconds, but it’s typically within seconds you lose this impression of having a real person in front of you. There is also no intelligence in the sense that these animatronics would stray too far away from the script.

Q: Can a machine act intelligently? Can it solve any problem a person can solve?

A: Definitely. Machines can solve problems humans cannot solve. For example, take the Rubik’s Cube, which for a long time was seen as a tricky task only smart people with dexterity and motor skills can solve. Nowadays robots are much better than humans at that. They can solve it in seconds. Recently I read a report where a robot solved it in a second. There are some tasks where robots can achieve super-human results, but I think what is so unique about human beings is that they are so adaptive and versatile. … We can do many, many things, and survive in a changing and challenging environment. Robots today typically can only perform a single task very well, even better than a human. But they are not able to transfer the knowledge on to new tasks or new environments.”

Q: What are your thoughts on creating an android of a relative, like a deceased grandparent or a child?

A: This of course is a very loaded question. I was working with a professor in Japan who created copies of himself and his daughter. These robots look very humanlike. Appearance-wise there is a quite a lot of progress we have done in the field, but again it’s a question of convincing results in behavior. This immediately brings the question of robotics and ethics to the surface. This is not something to take lightly and the community isn’t taking it lightly. There is an entire technical committee in the Society for Robotics and Autonomous Systems that is looking at the ethics of robotics. … There is an entire universe of ethical questions out there that need to be addressed. It’s not only the robotics community that has to think about that. It combines people from psychology, law — what does law tell us? What is the right decision from a legal point of view? At the moment I think the question of bringing deceased people back to life is too far away, it’s too Hollywood. You would not bring back the soul of that person.

Japanese roboticist Hiroshi Ishiguro created a copy of himself. Photo courtesy Heni Ben Amor.

 

Q: How would you feel if you found out that someone you were working with is a robot?

A: The question is: Does it make a difference? Ultimately, I think the way we perceive the world is through the behavior of a person … We think, ‘Today he is nice.’ We try to figure out what is his internal mental status, and that is how we perceive other people. Ultimately, if a device is convincing enough to make me simulate its internal state, then I don’t really care whether it’s human or robot.

Q: But you would want to know ahead of time?

A: I’m completely indifferent. I think this is also the case for many people. Many people have watched an animation show. In animation, all of what you see if fake. It’s not true. What the word “animation” stands for is “life.” It does not stand for movement. … The idea is to create this illusion of life, and people get attached to something that does not exist. Many people have cried at the end of an animated movie. … Any artificial system is too far away from our human capabilities at the moment. People may be impressed by what you see on TV, but it’s not much different than these animated characters to which you get attached. If you get attached to it, it doesn’t really matter whether it’s an animated character or a robot or it’s human. Believing these robots could get a soul and become conscious and sentient, that part is horrible, and we are not there. But if a robot does something funny, I’ll laugh at it. I don’t care whether it’s a robot or not.

Q: One thing that strikes me in talking with you is that when you study robotics, you’re studying humans in a way.

A: We are very much inspired by biomechanics in order to figure out how humans manipulate the world, how we walk in the world. It’s very similar to how aerospace engineers study birds in order to get inspiration and basic principles. Once you have those basic principles, you can create an engineered solution and follow the same approach. … Similar to that, we as roboticists look at humans in order to get inspiration about modularity. A human being is not one big piece; it’s not this monolithic thing. There are millions of pieces inside that work together in synchrony. Creating synchrony among millions of pieces is actually really tricky. Even the smallest mistake can completely offset the system and make it fail. Despite that, we human beings seem to work perfectly. If you look at the billions of neurons in your brain, they work in synchrony. They also work with the muscles. It’s also about the physical composition of our bodies and why our muscles are created in a specific way and how that is shaped by the environment. ... If I need a robot for a manufacturing environment, then its appearance and behavior needs to be completely different from a robot that roams around the desert to find water. We definitely study a lot of biomechanics and how human cognition works.

 

Top photo: Courtesy of HBO

Answer: The one on the left is an android created by Japanese roboticist Hiroshi Ishiguro. 

Scott Seckel

Reporter , ASU Now

480-727-4502

 
image title

Experts see steady economy in 2017 under Trump

Experts predict stable economy in 2017 under Trump at ASU forecasting event.
December 5, 2016

ASU economic forecaster predicts that Arizona will see more jobs added

Despite a wild and unpredictable campaign season, three top economists predict a stable — and potentially positive — economic outlook for 2017 under Donald Trump, who will be sworn in as president in January.

Unemployment and inflation will likely remain stable, and any policy changes wouldn't have big effects until 2018, according to experts at the annual Economic Forecast Luncheon on Monday, sponsored by the W. P Carey School of Business at Arizona State University and JPMorgan Chase and Co.

James Bullard, president and CEO of the Federal Reserve Bank of St. Louis, predicted that the most immediate economic changes could be seen with Trump’s promises to deregulate businesses, spend billions on infrastructure and reform taxes.

“If you think there’s been over-regulation of unwarranted regulations, you could see improvements in productivity, but regulation is a vast area that covers many different aspects of how businesses operate,” he said to the 700 people who attended the 53rd annual luncheon in Phoenix.

“The U.S. growthThe U.S. economy is growing at about 2 percent – a figure that Bullard predicts will hold for 2017. rate is low, but it could be influenced by those policy changes.”

But Bullard said that changes in trade agreements or immigration policy would likely take much longer to have an effect on the American economy.

He also said he was not concerned about remarks Trump made during the campaign that called into question the credibility of the U.S. Federal Reserve, which sets monetary policy.

“Trump’s transition team has said they’ll respect the Fed, and I take them at their word that ultimately they will endorse the current Fed structure and that we’ll be able to continue to deliver good monetary policy under a new administration,” he said.

Another economic expert predicted that if Trump decreases the corporate tax rate, as he’s promised, it could drastically increase profits — though not right away. Anthony Chan, the chief economist for JP Morgan Chase and Co., said the current corporate tax rate is 35 percent — although the average rate that corporations actually pay is closer to 27 percent.

“If the corporate tax rate is lowered to 15 percent, it has the potential to boost corporate profits by 19 to 20 percent,” he said. “But I’ll be the first to tell you that it won’t happen in 2017.”

Chan also said that Trump’s plan to allow corporations to repatriate profits made abroad could potentially raise $160 billion in revenues, which could pay for his infrastructure plan.

“As investors it’s not our job to say ‘this is good or this is bad.’ It’s our job to set our portfolios to benefit from these things,” he said.

Lee McPheters predicts more jobs for Arizona next year, but he worries about long-term indicators.

The outlook in Arizona is positive for next year — with some ominous long-term economic issues, according to Lee McPheters, research professor of economics in the W. P. Carey School of Business and director of the school's JPMorgan Chase Economic Outlook Center, which specializes in economic forecasts for Arizona and the Western states.

McPhetersMcPheters is editor of the Arizona Blue Chip Economic Forecast and the Western Blue Chip Economic Forecast newsletters, published monthly by the center. showed that of five economic indicator forecasts he made a year ago, three were better than predicted. There were 76,000 new jobs, compared with 68,000 predicted. The employment increase was more than expected — 2.9 percent increase in jobs compared with 2.6 percent predicted. And the unemployment rate was better than forecast — 5.2 percent compared with 5.8 percent.

The two indicators that were not better than McPheters predicted were population, which increased by 1.6 percent compared with the forecast 1.7 percent, and single-family housing permits, which increased 10 percent, not the 30 percent he forecast.

“If you look at Arizona’s numbers, we’re pretty certain as we wrap up 2016, we will definitely be in the top 10, and maybe the top 5, nationally for private job creation, and we expect that to continue in 2017,” he said.

Even with the positive projections, Arizona is below the national average in other measures of economic prosperity. The state ranks 42nd in per-capita income and 45th in poverty. The state also ranks last in per-student funding for universities.

“We need to look at policies that propel Arizona to look more like Colorado or our neighbors who have made the transition to using technology to increase income,” he said.

 

Top photo: James Bullard, president and CEO of the Federal Reserve Bank of St. Louis, predicted that the most immediate economic changes could be seen with Trump’s promises to deregulate businesses, spend billions on infrastructure and reform taxes. He spoke at the Economic Forecast Luncheon on Monday, sponsored by the W. P Carey School of Business at Arizona State University and JPMorgan Chase and Co. Photo by Charlie Leight/ASU Now.

Mary Beth Faller

reporter , ASU Now

480-727-4503