Position: Home > News > industry
IBM man machine debate 3: AI benefits more than harm narrowly wins, robot justifies itself
Core hints :An IBM robot recently entered a debate about whether AI will bring more disadvantages, and narrowly won the argument tha

Release date:2019-12-05

Browse number:18

 An IBM robot recently entered a debate about whether AI will bring more disadvantages, and narrowly won the argument that the advantages outweigh the disadvantages.
 
The debate took place at the Cambridge Union. The Cambridge union debating hall is 150 years old and has welcomed Margaret thatcher, franklin Roosevelt, Stephen Hawking, and even porn actors, from prime ministers to presidents, stars to leading academics. But never mind that 300 people will be watching today's debate between humans and non-humans.
 
IBM man machine debate 3: AI benefits more than harm narrowly wins, robot justifies itself
 
 
Man-machine debate "will AI bring more disadvantages" (source: IBM)
 
The robot, a conversational AI from IBM's Project Debater, is a two-metre-high black box that sounds like a female celebrity. The debate began with his opening remarks. The robot has a 'speech by crowd' app that analyzes comments submitted online to generate its own arguments. It relies on more than 1,100 pre-submitted human statements.
 
Two debate teams are composed of a professor and a debate expert, and both teams are assisted by the robot's AI. "I'm really curious and excited to work with a machine." "Sharmila Parmanand, one of the participants and a debate expert for the anti-ai team, said before the event.
 
"Ai cannot make morally correct decisions because morality is uniquely human." There was silence. Normally the audience's attention would have been focused on the union President, who was the highest in the auditorium, but now all eyes are on the debater, who is flashing blue.
 
"Ai has a lower error rate than humans... That would be a huge advantage because it would save more time to perform mundane and repetitive tasks." It raises two opposing points of view.
 
Proponents of AI argue that AI will create new jobs in certain fields and "increase productivity in the workplace."
 
The robot then goes on to suggest that "artificial intelligence can take care of patients, or robots can teach students -- areas where humans are no longer needed."
 
Ai had a coherent debate, but there were also some missteps. Sometimes it repeats its ideas, and sometimes it fails to provide detailed material to support its ideas.
 
It's an interesting debate, but it's not just a debate. The technology IBM demonstrated this time is very different from previous demonstrations.
 
IBM man machine debate 3: AI benefits more than harm narrowly wins, robot justifies itself
 
 
In 2012, IBM Research established a Project team and officially launched the development of "Project Debater" (source: IBM Research).
 
The original proposal for the debater project grew out of a single powerpoint slide that appeared in 2011. Because of the limited submissions and the difficulty of challenging the project, it took about a year of careful consideration before IBM management finally approved the project. In 2012, IBM research established a Project team to officially launch Project Debater. Substantive work on the project did not begin until 2014.
 
In June last year, IBM's debate bot held its first debate with a human, scoring 1:1, with viewers voting that Debater was less skilled than a human but more informative. At the time, the system had a library of hundreds of millions of articles from which to extract content on about 100 topics. When confronted with a debatable question, it takes a few minutes to search the articles, build the argument, and make the overall presentation.
 
In February, at IBM's Think conference, AI robots revisited champion debater Harish Natarajan. AI lost a bitter battle over whether the government should fund pre-school education.
 
In the past, this kind of argument with people mainly trained AI's listening comprehension ability and ability to make a meaningful response to an argument. These skills may be far beyond the capabilities of smart assistants in smart homes or mobile devices, but they fall far short of what IBM was trying to do in the first place.
 
This time, it's in the form of an AI self-argument, as if "you put together thousands of positive and negative restaurant reviews and summarize them in a few paragraphs," said IBM computer scientist Dan McCullair. "He said.
 
So the point of this debate is not whether machines can persuade people, but that machines can enhance them.
 
"Humans can better understand our audience based on our culture and traditions, but machines can enrich our knowledge with vast amounts of data to make more specific decisions that allow us to eliminate the biases inherent in being human." Lahav said.
 
Machines find valid connections between things that are likely to be ignored by default. This limits our ability to process information.
 
"In 2016, Debater was just a toddler, and by 2019, we're already at the college level," said noam romer, the lead researcher on the project and a distinguished IBM engineer. "Project Debater can help humans in many ways by showing both sides of a controversial topic and reducing inherent human biases."
 
Debate has long been an intellectually challenging task. Getting humans to understand an argument and say yes or no is simple for us, but incredibly complicated for machines. The next challenge for the robot is to determine whether the quality of the argument is sufficient to convince others. Finally, ai must make its case in a narrative that humans can understand.
 
In addition to debating, the technique could be used in any job that requires collecting data and making decisions based on it. "Think of it as a research engine, not a search engine," Lahav says. Lawmakers can see each other better because the machine can debate itself."
 
So far, the "debater" has done a good job of organizing the language in a way that doesn't put the audience to sleep with the sound of a machine in the chilly November season. Instead, it makes the debate light and lively.
 
 
At the Cambridge union, people "vote with their feet" (source: Medium)
 
Sylvie Delacroix, a professor of law and ethics at the University of Birmigham and a member of the opposition, took to the podium and jokingly referred to Project Debater as "Debbie." "Ai is an important tool because it is changing the way we work and the way we date at an absolute rate." She said.
 
Instead of voting immediately after the debate, the crowd stood and cheered. After that, the audience walked to the three doors representing "yes", "no" and "abstention" respectively. The final voting result was 48.17% in favor, 51.22% against and 0.61% abstained.
 
"I really don't know what the future of ai is," Natarajan said. "it's definitely going to have a huge impact, but we don't know what it's going to be."
 

Friendship link    | International Artificial intelligence

News Exhibition Video Product Enterprise Picture Technology
Liaoning ICP No. 234234234 -1 | Liaoning No. 1102342342030919 public security