Stay updated with the latest in Tech, Science, Culture, Entertainment, and more by following our Telegram channel here.
There are plenty of real-world applications for artificial intelligence (A.I.), but who’d have thought that arguing with humans would be one of them?
A new A.I. program developed by IBM is apparently getting really good at debating against humans that it’ll probably be impossible for regular people to win arguments against it in the near future.
Project Debater being publicly demoed in 2019. IMAGE: IBM Research
The program is called Project Debater, and has been in development for the past several years. Its sole purpose is to argue against humans on an array of topics in a meaningful and constructive manner, meaning that it’ll take sides in debates and look to make strong logical arguments for its chosen viewpoints.
In 2019, the program was given a public demonstration where it entered into a debate against expert debate champion Harish Natarajan on whether pre-school education should be subsidized for families.
Although Project Debater lost against Natarajan in during the exhibit, the program proved its capability to form logical arguments on deep topics and present strong arguments to support its stand.
Lots of data required.
Project Debater does all of this via a technique called argument mining. This technique involves the A.I. sifting through vast amounts of information relating to the chosen topic, and then linking relevant portions of data together to form reasoning and supporting arguments.
For Project Debater to perform argument mining, it consumes hundred of millions of news articles and uses the information derived to formulate opening statements, propositions, counterarguments, and closing summaries regarding over 100 subjects.
But while its huge bank of information will probably cause some concern amongst debaters coming up against it, Project Debater still has some way to go before it can pose a serious challenge to today’s human debate experts.
However, observers given the task of rating the program’s arguments have scored them quite highly, and have marked Project Debater’s performances as decent in the very least – quite impressive considering that such a feat was considered almost impossible just a decade ago.
“A combination of technical advances in A.I. and increasing maturity in the engineering of argument technology, coupled with intense commercial demand, has led to rapid expansion of the field,” said Chris Reed, a researcher commenting on the project.
In 2017, an A.I. bot called Open AI convincingly beat some of the world’s best Dota 2 players at their own game. IMAGE: Shacknews
Recent times have seen A.I. prove its superiority over humans in a number of competitive arenas, but the ability to form strong and convincing arguments during a debate is considerably more challenging. Thusly, this means that the task of using argument mining to win debates isn’t as straightforward as, say, beating an opponent in a video game.
“Debating represents a primary cognitive activity of the human mind, requiring the simultaneous application of a wide arsenal of language understanding and language generation capabilities, many of which have only been partially studied from a computational perspective as separate tasks, and certainly not in a holistic manner,” said a study done on the subject.
“Therefore, an autonomous debating system seems to lie beyond the reach of previous language research endeavors.”
What does this all imply, then? Well, it means that even if A.I. programs like Project Debater are getting way better at piecing together data for tasks such as debates, we shouldn’t worry about them surpassing us in cognitive reasoning and argumentative prowess anytime soon – although such incidents could happen further down the line.
Read more tech stories:
Cover image sourced from CNET.
ที่มา : Mashable