Science Fact: The end of paralysis is the promise, a computer takeover the threat. But the news features sniping by scientists. Who’s in the right, anyway?
This is a timely post, because three hours from now there may be a noteworthy public demonstration of paralysis-fighting technology.
Think about these two headlines:
– Kickoff looms for demo of brain-controlled machine
– Thinking machines are ripe for a world takeover
The first of these articles describes a machine developed by Brazilian neuroscientist Miguel Nicolelis to help a paralyzed person walk. It’s a demonstration planned for today at the opening ceremony of the World Cup in São Paulo. It relies on a framework called an “exoskeleton” – think of metal braces strapped around the legs – and a skullcap that senses brain signals without electrodes penetrating the skin. The user thinks the command and the exoskeleton causes him to stand up, walk forward, and since this is the World Cup, kick a soccer ball. A triumph of technology over physical disability!
The second article describes a piece of software that pretends to be a person. The software was developed in Saint Petersburg, Russia and simulates a thirteen-year-old boy from Ukraine whom they call “Eugene Goostman.” Kevin Warwick, a visiting professor at University of Reading, UK, organized a test in which thirty judges tried to tell whether “Eugene” was a normal weird teenager or a computer program. A third of the judges thought he was human! And the Financial Times’ take on this demonstration? Machines are about to take over the world!
There you have it: the greatest promise and the greatest threat from computer technology, simultaneously thrust in front of our faces. Polar opposites.
Both these advances deserve a few more sentences, but first it’s worth noting the way in which these two stories are exactly alike: they both involve scientists pooh-poohing the work of other scientists.
Nicolelis is criticized for demonstrating what’s basically a robotic prosthesis not between the pages of a musty journal, but in front of 62,000 fans in Arena Corinthians and hundreds of millions more on TV. Nicolelis seems to take delight into goading his critics. He says he’s planning a demonstration for the Olympics, which Brazil will host in 2016, but he won’t give the details yet: “we’ll tell them in due time…I’ll let the critics have one more month of good sleep.”
Kevin Warwick’s demonstration has attracted similar sniping. For example, an NYU professor says that the test “proves the power of fakery.” The complaints seem to center on the fact that the computer is pretending to be a young boy with poor command of English, and the judges only had five minutes of conversation to make a decision, so on balance nothing new was demonstrated. But there’s another angle too: a sour blogger refers to Warwick as a “serial hoaxer.”
The underlying similarity is evident: step out of the lab and into the spotlight and you will easily be able to fool journalists with your technical mumbo-jumbo. However, your scientific colleagues, driven by a mixture of envy and intolerance, will hate you for presuming to be a public figure who speaks for the profession.
It’s no wonder that so few scientists succeed in that delicate balance between being part of their field yet also stepping outside it for the media.
Science Speculation: Let’s set aside the politics of media scientists and return to the work. Is there something here of importance?
Where the paralysis-fighting prosthetic is concerned, I think yes. Nicolelis may have his eye on keeping his research funding rolling in, but a dramatic demonstration of robotic assistance is likely to raise awareness and support for similar work across the globe. And there are special features to Nicolelis’ setup worth noting: direct brain control without electrodes; high-level programming for the motions; and feedback signals from the prosthesis through a special shirt that the patient wears. Nicolelis asserts that this noninvasive, two-way communication helps the users feel as if they are actually performing the motion themselves. Given the brain’s amazing plasticity – ability to adapt and re-wire itself – that sounds not only plausible but tremendously empowering. Not only the end of paralysis, but an end to feeling paralyzed. And that qualifies as very important.
Where “Eugene Goostman” is concerned I reach the opposite conclusion. In 1950 mathematician Alan Turing proposed in a famous paper that since we don’t really know what “thinking” is, rather than ask whether a computer can think, we should determine whether it can convincingly imitate a human being in conversation. This has come to be called the Turing Test. Warwick captured headlines by claiming that this was the first time a computer had passed the test.
Alas, there are problems with his claim. For one thing, there are many versions of the Turing Test and Warwick’s criterion – that 30% of the judges be fooled – does not seem very demanding. In addition, “Eugene Goostman” was introduced as a young teen from a foreign country struggling to communicate in English.
Those who have read Death By Probability will recall that in creating his artificial intelligence friend Al, Evan was inspired by Cleverbot, a program that converses by feeding back to you things that have been said to it in the past. And Cleverbot has passed a number of Turing Test competitions: in one of them, in 2011, it fooled 59% of the judges.
It used to be that anyone could converse with “Eugene Goostman” on a website (but, as of August 2015, no longer). I tried it myself and found the program to be unconvincing. It asked questions about me, perhaps so it could tailor its answers, but the questions were quite repetitive. It also said some strange things, such as “In my native Ukraine people do not normally eat.” You might also like to read an actual conversation that blogger Scott Aaronson had with “Eugene.”
What can we learn from these examples? Things that we already knew, such as:
– We can’t trust media to separate hype from real science
– We can’t automatically dismiss the value of scientific work just because it attracts criticism from other scientists.
However, we (or I, at least) also learned something new – that despite its self-promotional overtones, something important is occurring today in Brazil: a courageous public demonstration of work that may signal, in many ways that count, an end to paralysis.
I don’t know about you but if my cable service cooperates, at 2:00 pm Eastern today I’ll be watching the Opening Ceremonies of the World Cup, rooting not for Brazil or Croatia, but for Miguel Nicolelis!
News Update Added Late on 6/12/2014:
The Miguel Nicolelis soccer-kick demonstration, as part of his Walk Again project, was successfully executed at the World Cup Opening Ceremonies this afternoon. Volunteer Juliano Pinto, 29 years old and paralyzed in his lower trunk, wore the exoskeleton. Commanding the prosthesis with his brain waves, he kicked a soccer ball a short distance to symbolically start the 2014 World Cup competition. The event was covered on some but not all TV broadcasts of the Opening Ceremonies and some observers felt it did not get all the coverage it deserved.
The media documentation is here:
– BBC report of the symbolic kick: Paraplegic in robotic suit kicks off World Cup
– YouTube video with the kick slightly after the 1 minute mark: World Cup 2014: First kick made by Paralyzed Man with the help of mind-controlled Robot
Paralysis. Human-like computers. When scientists can’t agree on the significance of scientific work, how is the general public supposed to make sense of it?
Drawing Credit: Adapted from kolbasun, on openclipart.org
I have updated the original post with a blockquote paragraph near the end, reporting on the demonstration conducted at the World Cup Opening Ceremony.
Charles, your comments are always interesting to read. In this case pretty funny too, since you cause me to wonder whether a computer acting human is more or less noteworthy than a scientist acting human!
Powered exoskeleton “suits” are of course an active interest of the military, given their desire to enhance abilities of soldiers on the ground. In addition to the probably greater strength and endurance of such a suit, it allows armies to clad the suits with extensive armor to protect against projectiles and explosives that can save lives of soldiers on the ground. The military’s interest in such technology almost assures that the breakthroughs will eventually work their way down to those people who are paralyzed or handicapped.
As for the “chatterbots” (including Cleverbot) … my first experience with that category of software application was in 1980 when HRL got their first VAX. The standard software included a program called ELIZA, one of the very first implementations of software designed to converse with a human. It was entertaining to see the responses of the various scientists at HRL when they first launched the program, as it drew them into a give-and-take conversation that for the first few minutes seemed very humanlike.