The Use of Artificial Intelligence (AI) to Help Reduce Conflict and Misunderstandings in Meetings
Introduction
The concepts outlined in this article are an attempt to stretch the imagination into what can be possible with Artificial Intelligence (AI). In this instance, I will be considering the possibility of an AI program that can monitor the mood and communication styles of people that meet regularly such as a board or committee.
What is AI?
First of all, perhaps a definition to be considered in the context of the article. AI, as anyone would probably know by now is a branch of computer science. The application results in the development of machines that are capable of performing many tasks through the use of algorithms (Al-Fedaghi 2018), or set of instructions, that are normally the domain of humans, but potentially at much greater speeds. Some of these abilities could include visual perception and recognition of speech, even translation.
AI performs these tasks by referencing what are often referred to as Large Language Models (LLM). These are exceptionally large datasets of information that are used in generating responses to prompts, inputs or queries.
How AI can be used to reduce conflict and misunderstandings in meetings?
This article is focused on the concept of small groups given the technology being discussed is not yet available. It is far more simple to describe expectations and outcomes for a small group than a large diverse one.
When it comes to people, there are many personality tests that can be done for defining dominant styles of behaviour and communication. Some of these models will be considered further on. Most however, put people into one of four quadrants or classifications and attribute certain communication and behaviour styles to these. It should be noted however that everyone is a mixture of all styles. There will, however, be a dominant style people will revert to as their default style, particularly when under stress or most relaxed. It is beyond the scope of this article to go into personality and communication styles, however, the reader should be aware of the impact they can have on a group discussion and dynamics, indeed, even one of an individual nature.
To assist an AI model in identifying mood, attitudes and communication styles, it would be helpful if members of a group completed a personality survey and fed the results into the model. In addition, the model would also require the ability to recognise the voices of the people speaking in order to match the voice to the personality style.
Consider the differences between the personality types in the image below.
Figure 1: Personality types by colour
Given such basic information as above, people themselves could be much more self aware of themselves and their behaviour in meetings. Generally, this is not the case, and we respond in accordance with our basic drives. What is being suggested here is a scenario in which an AI model could be “reminding” people of the dynamics in the room.
It could not be expected that such technology would be welcomed with open arms by everyone and it is understandable if people were to treat it with a degree of suspicion. This issue will be addressed further in the article.
Benefits of Using AI in Meetings
Given the above basic overview, we can now turn to the benefits that might be achieved through the use of AI in meetings.
Meeting Records
Although not really new, a model developed along the lines of those proposed, should be able to record meetings and provide transcripts for minuting purposes. Hand written minutes always rely on the skill of the minute taker and the interpretation they place on the conversations which may not always be accurate or complete.
Flags on communication or behaviour of individuals could also be used for training of the people involved. I must stress, however, that I would not like to see any form of AI involving people used for “policing” purposes in an organisation. It should always be used to empower people and help them grow to reach their maximum potential. Such growth is of benefit to both the individual and the organisation.
Improved Communication
As an example, let us take two ends of the communication spectrum. Some people can be quite verbose in speech while others can be short and sharp, perhaps even blunt. Can you imagine the friction between these two communication styles? One could feel short-changed on information while the other could be bored with the length of time it is taking to explain something.
As an attendee at a meeting, the AI model could be continually scanning the environment and providing you with feedback. No doubt you would have your own ideas on what might be going on, but the model could potentially either reinforce your views or provide an alternate possibility.
Communication is a huge issue at all times. This is compounded by the well established fact that the majority of our communication is non verbal (University of Texas, n.d.; London Image Institute, 2020; Wilson 2021). Figures vary widely, but there is no doubt that it does form a considerably large part of our communication process. The potential for an AI model to get real time feedback on this via cameras in a room is certainly a possibility, but perhaps that is too far into the future. It also serves the purpose of this article to concentrate on the spoken or written word.
A favourite saying is “that the meanings of the words are not in the words, but in the people”. So, if I was to use a particular word, it may have certain connotations for me, but could potentially be a trigger for you because of the meaning you attach to the word.
As an example of this, I can recall attending a meeting once where someone proposed that some volunteers were going to form a sub-committee to undertake a project. What this person actually meant was that a small, informal group of people were going to get together and undertake a project with benefits to an organisation, but not reporting to it in an official capacity. They didn’t mean a sub-committee at all in the sense that it would “report” back to the Board.
Other people listening took the term “sub-committee” and applied it in its correct context that it would be a committee that reported back to the Board and, therefore, a part of the formal structure of the organisation. This was not the case at all, and conflict arose over the simple use of a word and the misunderstanding that arose. In hindsight, perhaps the best course of action would have been to ask the person who raised the issue what they meant by a sub-committee. The opportunity was missed to avert the conflict and other people reacted based on their meaning of the word. This type of thing goes on all the time, as I am sure you will have witnessed.
So, the question is whether an AI model could listen for these types of misunderstandings and flag them before they escalated? Could the AI model perhaps even suggest that the person be asked what they actually meant by a word they had used?
In reality, this comes back to the age-old problem of asking someone to complete a task. Your idea and their idea of “done” or complete could be totally different. It is always worth agreeing on what done looks like to both parties. Words are no different. We need to have all parties agree on the meaning of the words in the context they are being used.
Increased Efficiency
Any tool that can assist in cutting through communication problems can lead to increased efficiency in the running of meetings. Such efficiency could improve in several areas, but there are two notable ones that immediately present themselves.
First, through more effective communication, the ability to process language could be improved. With an improvement in the speed of language processing, the quality of responses and feedback should also be improved and conflict minimised or at least mitigated. Secondly, it may therefore be possible to reduce the amount of time spent in meetings due to the better communication processes. The less time spent in meetings, the more time that is available for other productive pursuits and of course these are not limited to work pursuits either.
Potential Challenges of Using AI in Meetings
The largest of these would initially be suspicion and resistance to change. New technology has to be “sold” and the only way to do that is to allow people to prove to themselves that it works. No amount of sales spiels will achieve this.
Let us take an example of a hammer. If you were told that there is a brand new tool for driving nails called a hammer, you may well be sceptical because you have always used a rock in the past. The hammer may be demonstrated to you, but you are happy with your special rock and see no need to change. It will only be once you use the hammer yourself to drive several nails and realise its superiority over your rock that the rock will be retired.
New technology in the AI space is no different. It will be treated with scepticism and uncertainty. It is a fear of the unknown. Here would be an excellent point in the article to refer to the Product Adoption Curve or Diffusion of Innovation Theory developed by Everett M Rogers (Lambert 2018). This theory demonstrates how products, or in this case, technology, is adopted. It is not a simple matter of introducing something new and adoption is 100%. Anything new requires patience and support in adoption.
Figure 2 — Diffusion of Innovation Image
Again, this is another area beyond the scope of this article, but awareness needs to exist around technology adoption to understand the reluctance of people to change and the need for patience.
Potential for Misuse
Any tool, irrespective of whether it is physical, mental or digital has a potential for misuse. Artificial intelligence is no different. The two most significant areas in the space are the questions of ethics and bias.
In our proposed case, the building of the model around personality types and traits would be critical to reliable feedback. The model used in this article for example purposes has not been based on empirical studies, but on practical experience. It is questionable as to whether or not this is a reliable path to go down when developing such a proposed model. I would contend that any dataset needs to be based in solid research and also keep up to date with the ever changing nature of these studies.
Many people would be familiar with the Myers-Briggs Personality Test (The Myers & Briggs Foundation n.d.)and DiSC theory (DISC: The History of DISC Personality Styles — DISC Insights n.d.). The links, which are in the references, are provided purely for further reading. Another set I have seen in the past are People Focused, Detail Focused, Action Focused and Innovation Focused. The model you prefer will be up to you or your organisation. There are certainly several to choose from.
If the datasets are flawed in any way, and it is arguable that they are all flawed, it is just the matter of degree, then there is potential for bias. There are many credible articles online about bias in AI datasets and I would encourage you to explore them to become aware of the problems.
Bias, of course, can lead to prejudice and if our imaginary program is spitting out biased information, and people rely on that, then they will be misled. The purpose of the program imagined is one of an assistant, not one of a master!
The ethics around using such a program are murky as ethics often tend to be. A significant potential ethical misuse though could be using the program to argue the incompetence or standing of others based on the output. As an example, if you have someone who is somewhat dominant in a group and the model is telling them that the person currently speaking is anxious or appears confused, then there are two ways to use such information. The program output could be used to humiliate someone, or it could be used to empower them through loving feedback and assistance. If the former method is used, then the model should red flag that behaviour and the person using it. If the latter behaviour is used, then kudos to the person delivering the support.
The bottom line here is that, like any tool, AI can be abused or used for the communal good.
Conclusion
Because of my personal involvement in many organisations over the years, whether they are community based or business based, I have seen misunderstandings and conflict occur frequently. Any opportunity to mitigate such problems should lead to improved efficiencies, less conflict, more harmonious working relationships and great empathy towards others. Perhaps it is a utopian dream, but a world where people all get along and understand each other is certainly a virtuous goal to work towards.
THANK YOU!
Want access to the best stories and articles on the web? Subscribe to Medium through my referral link for just $5 a month. You’ll have access to exclusive content and valuable insights that non-members miss out on. Don’t wait, elevate your reading experience and support me by subscribing now here.
References
Al-Fedaghi, S 2018, ‘Exploration in Algorithm Engineering: Modeling Algorithms’.
Aydoğan, R, Baarslag, T & Gerding, E 2021, ‘Artificial Intelligence Techniques for Conflict Resolution’, Group Decision and Negotiation, vol. 30, no. 4, pp. 879–883.
DISC: The History of DISC Personality Styles — DISC Insights n.d., viewed 18 January 2023, <https://discinsights.com/disc-history>.
Gregorc, AF 1986, Adults Guide to Style, Gabriel Systems.
Lambert, P 2018, ‘Innovation Diffusion: A Complex Adaptive Process’.
London Image Institute 2020, How Much of Communication is Really Nonverbal?, PGi Blog, viewed 14 January 2023, <https://www.pgi.com/blog/2020/03/how-much-of-communication-is-really-nonverbal/>.
The Myers & Briggs Foundation n.d., viewed 18 January 2023, <https://www.myersbriggs.org/>.
Ulrich-Tobias, C 1999, The Way We Work: What You Know about Working Styles Can Increase Your Efficiency, Productivity and Job Satisfaction, B&H Publishing Group.
University of Texas n.d., How Much of Communication Is Nonverbal? | UT Permian Basin Online, viewed 14 January 2023, <https://online.utpb.edu/about-us/articles/communication/how-much-of-communication-is-nonverbal/>.
Wilson, CR 2021, Nonverbal Communication Skills: 19 Theories & Findings, PositivePsychology.com, viewed 14 January 2023, <https://positivepsychology.com/nonverbal-communication/>.
More content at PlainEnglish.io. Sign up for our free weekly newsletter. Follow us on Twitter, LinkedIn, YouTube, and Discord.
Interested in scaling your software startup? Check out Circuit.
