Why an AI System to Debate Humans?
IBM are developing the Project Debater, with the use of AI, to allow computers the ability to create points of view across a range of topics and subjects. It has been 6 years in development and the system is being designed to have meaningful discussions with humans where with the use of artificial intelligence it can research quickly and develop a range of ideas to discuss and offer counter arguments with humans.
The idea is for computers to have meaningful discussion bringing value. IBM are looking foe their Project Debater machine to bring a range of views with deeper ideas to various topics and maybe bring about thinking that hasn’t been reached before.
The system is a 6ft tall black box have even given it a screen for a face and a mouth. In fact, IBM have actually run some live debates and we have included a YouTube videos of the second debate below. This has some amusing statements made by the Debater system, see for yourself below.
The project has been in development since 2012 and follows on from the IBM Watson project.
Here IBM discuss the background and thinking behind Project Debater –
IBM Project Debater – First Live Debates
Arvind Krishna IBM announced the first ever live debate with humans on the 18th June 2018. Two debates took place to allow parties to consider different arguments and to hear and consider opposing and differing views. The interest being on how computers and artificial intelligence can look to bring points of view that may not yet have been considered by humans.
The 1st debate was on whether there should be more publicly funded space exploration. The 2nd debate was on whether more should be invested in telemedicine with each participant had 4 minutes to make their opening statement, followed by a 4 minute rebuttal and then finally a 2 minute conclusion.
The machine considered and rebutted a number of arguments drawn from a library or millions of documents, including academic journals to form it’s opinions and put forward it’s arguments. IBM say that the specific chosen debate topic was on subjects not given to the machine beforehand though a 100 or so subjects were given to the machine beforehand that were thought to be potentially covered in the debates.
The human audience were given a vote as to whether IBM Project Debater or the human had won the debate. The members of the audience widely agreed that for both debates the humans, Noa Ovadia in the 1st and then Dan Zafrir in the 2nd, who are both Israel national debate champions, made greater delivery with the audience agreeing that the IBM machine had greater delivery.
Below the YouTube video shows a Fox News interview with IBM Director of Research Arvind Khrishna –
IBM Project Debater – What does the future hold?
So what can we expect in the coming years as the AI machine learning develops –
- The machine is being proposed in the future to be utilized in the corporate boardroom where there are biases and differing opinions to put forward clear concise arguments for the good of the company.
- It could be used to debate our politicians and to offer a deeper argument for and against decisions to assess their implications.
- There is also talk of this kind of technology being used in the courtroom as your lawyer once the technology develops sufficiently or more likely in the first instance to form and tweak arguments during the hearing.
There will certainly be greater use of this kind of technology across a number of areas where humans need to debate and assess different ideas. Where else do you think this kind of AI machine could be utilized?