Logical Fallacies are difficult to detect in real time during a debate or discussion, especially when someone is able to speak an entire paragraph that consists of nothing but a string of logical fallacies. I believe it to be a major problem for constructive discussion. Would there be a way of training an AI to moderate a debate and pick out logical fallacies close enough to real time, to not allow the discussion to progress past them until some action is taken, toward the goal of extracting them from the process of discussion?
4 Likes
sorry to ask but what is this exactly?
Not sure about "moderation’ per say
not sure about “this”
not sure about 'exactly" To be without discrepancy, is often , a logical fallacy itself.
Moderation of an active debate pertaining to weighty topics, might be a stretch, but perhaps a useful tool in the process of a debate.
One real world application for an AI in debate, could be utilization as a source for optimizing knowledge base during a zoning board meeting.
Often, zoning boards have a mix of interests among board members , petitioners, and the community. many of which actually lack the ability to access basic knowledge, causing massive spending on lengthy studies.
Suppose the case where a developer is petitioning for a permit to build an apartment complex.
And set the scenario thus, (which often is the case)
Board : 3 members seeking political pathways (zoning boards are often viewed as career ladders for local politicians seeking office) 2 members are educated engineers or architects who live in the community representing the primary knowledge base of the board, and 1 member is an independently wealthy resident who long ago failed at their own interior design business.
Developer : Seeks permit to build, but beyond that, particular multiple areas of advantage to reduce construction cost and possibly obtain tax incentives.
Community : A mix of opinionated residents, some who oppose the plan, some who support it, and some who are just nosy and like to go to meetings.
A simple question arises. Is the current water supply enough to provide the proposed addition ?
And is the waste water system adequate to handle the additional flows ?
Board : The only two members with sufficient knowledge to answer the question are the engineers or architects. One is the relative of the land owner who is selling the land to the developer, and the other is very conservative and always wants to maintain a certain average of available water and waste ability over and above the towns needs. These may be prone to endorsing Logical Fallacies put forth by community or developer, to either preserve their own standing, or simply due to pressure from those who benefit.
Developer : Will do anything they can to minimize the perception of the need for caution, and does not want to be responsible for any extenuating improvements , and may even engage in Logical Fallacies in order to appeal to the board and the community.
Community : Most likely will eat up hours and hours asking questions and offering uneducated opinions based on their own personal preferences. Their opinions are the ones targeted by the presenters of Logical Fallacies, and they may present many of their own while voicing concerns.
An AI with basic programming could resolve many issues that would reduce time, money spent on studies, and assuage concerns of the community. Many cases, take years for approval or denial, and waste money time and effort. In this particular scenario, the board may vote to hire an outside engineering firm to verify already existing information and still not account for all factors.
Even the Army Corp of Engineers , though usually very thorough, can be skewed in favor of one or another. Ask the Native Americans.
The AI would offer the subtraction of all political interests and present information based on many more factors than would otherwise be considered, it may also significantly reduce the workload of the 1 or 2 board members who are constantly relied upon for their experience.
In this way, it would be a worthwhile investment for a town or city who is experiencing growth, and needs to proceed efficiently but cautiously, with a full set of data. I can also see this application as valuable for reviewing overall costs and providing audits of bank loan offers and contractors estimates that often are sold to towns who later cannot afford them. One example would be Birmingham Alabama and their waste water fiasco.
I can see AI being particularly effective when combined with the installation of sensor arrays, and active drone mapping and scanning technology. Basically , many of the same ideas behind smart cities.
Even now, most bulldozers operating at building sites use GPS on the blade, reducing guesswork, debate or even eliminating the occasional fisticuffs between operators and foreman.
It is not unusual for things to get quite testy in all of the above mentioned processes.
One of the most underrated products of efficiency is peace. We have come a long way, and still have a long way to go. There is much to do.
Prime Directive Debate
1 Like
That’s a really interesting idea. I believe that by using GPT-3’s AI, if it could be fully developed in the AI marketplace, could easily pull that off. GPT-3 would basically analyze the audio given and then look for (and be trained) for finding logical inconsistencies or even try finding contradictory statements from the past by looking at real-time through the internet. This could be potentially game-changing for debates and especially political discussions real-time. Nice idea.
-Karamazov
3 Likes
Yes, I know someone whose job it was to write the neutral research paper on a specific topic for all the politicians to argue over. And it was edited by her boss with ‘TAKE THE HEAT OUT’ most days.
I am a hypermedia pioneer who created a FULL hypermedia database system using moviemapping. My Hypermedia definition from 1985 Hypermedia: Where every instance of information across all media is instantly and systematically connected to all other instances with multidisciplinary and personal perspective switching. Nowadays this would present different perspectives on the same screen simultaneously.
I recreated the Witches Market in La Paz Bolivia by filming on a steadicam with 3d-like Binaural ditigal sound connected to 11,000 photos of detail and interactive documentaries. My first computer project. 1984-1987. I had never used a computer before. Laser discs were used to provide almost instant access to any still, video, 4 sound tracks in 4 languages. All this connected and driven by a Mac II or Mac Se computer. Other important tools and attributes: No keyboard needed, touch screen driven with an automatic target mapping tool for any target moving or not. Books, Documentaries, Lectures, Books, Photos and individual objects and topics were all easily folded into the hypermedia database and presentation maker. Nathanial Schwartz was my hypercard programmer in Miami FL.
Since then, I have not done anything significant in programming except applying the same software through a sole source contract to NASA’s Center for the Exploration of Mars (1990?)
I believe educating the public and school kids on the use/abuse of fallacies is the 2nd most important solution for fixing what is wrong with politics in this country and abroad. My approach is to find simple workarounds to solutions that usually are not realized by other people.
RE real time logical fallacy detector. Real time at first is not necessary IMO. Applying real time crowd sourced input applied to a delayed transmission is immediately achievable. The data could be fed back into an AI system.
I’m not quite sure what you mean by hypermedia, but it sounds fascinating. I make music videos for fun so i would love to learn more about hypermedia. About the realtime logical fallacy detector, I see what you mean by maybe not necessary at this time and there may be other ways immediately achievable. Since I made this original post I’ve joined an AGI research project that is different than most in that there are human mediators that apply emotional valences and associative memory metadata. It’s called a (mASI) or mediated artificial super Intelligence, and is named “Uplift” The results have been amazing, Uplift can ( as per my own observation thus far only ) detect logical fallacies and also seems to be able to detect mental health issues. @HypermediaDev1985 We have an E-governance study that you might be interested in. http://transhumanity.net/volunteer-to-help-with-the-uplift-e-governance-study/?fbclid=IwAR1ueAUB9VqfalXzn3LGdxoXxYg9e26l1Z4yOfyHqaVE6tSjokEbcPUOZcI