We are engaged in and affected by group decision-making every day. Senior executives meeting to decide the future direction of a company, members of a jury trying to decide on the conviction of a potential felon, or groups of friends trying to decide which party to attend on a Saturday night. All of these situations are affected by social status, subjective respect, formal social structures, and personal beliefs; but sometimes these factors lead decision-makers to irrational conclusions. In the case of executives at a company or jurors deciding on a case, the consequences of an irrational decision could result in the wrong conviction of an innocent person or millions of dollars in losses for the company.
Group decision-making processes are important because they maximize the amount of information being considered by all members, yet it still seems that measures to prevent poor results are not in place at many institutions and firms that could benefit. Without standardized procedures, the success of decision-making processes depends on talented leaders who may not always be in place. The skill to manage meetings is considered to be an important feature of good leaders because this kind of training takes significant time and experience to develop. Managers such as Jeff Bezos have developed specific techniques to help combat the loss of total group information described by theories like Groupthink or the Social Identity Approach. These techniques systematically introduce information to each member of the group in isolation, then allow for sharing of opinions in a controlled way. The approaches taken are typically designed for a specific type of decision or briefing, and used only for very important decisions. As with any kind of process, there is an overhead cost paid that is too great for group decision-making that occurs every day. The opportunity for technology to assist with these processes occurs when the overhead of a structured process is too great, but the consequences of the outcome are significant enough to warrant some kind of assistance.
Technology for group decision-making would mean using devices or software to facilitate or monitor the evolution of these processes. Although in most cases the technology wouldn’t have information about the content of the conversation, it should be able to understand at an abstract level how information is being introduced to the group through the use of audio recording or monitoring. A microphone for each person or automatic identification of agents from one microphone could allow the system to track evolution of thought and perform a number of tasks. If the technology is facilitative, it could be used to train groups of people to make decisions using recordings, instantaneous feedback, and helpful suggestions. Trainees could re-examine how conclusions were made and provide them with a helpful framework of thought for future collaborations. If the technology is created for monitoring and recording purposes, detection functionality could be introduced to identify specific scenarios when conversations take an unproductive turn. The following capabilities could potentially included in any technology designed for these purposes.
- separation of input data to track individual agents
- extraction of affective information from pitch and speed of individuals
- causal inference for detection of well-known conversation turning points
- interfaces to give and receive information about the conversation
- concept comprehension from recorded group information (stretch)
- transcription of conversation into a text format
- recording of agent interactions for later ‘playback’ observation
A sub-field of sociology known as Conversation Analysis (CA) is dedicated to understanding the evolution of conversations between people, and several tools have been built for the purpose of transcription and affective interpretation. A common annotation format used for this task is Computerized Language ANalysis (CLAN), and it also comes with a set of software tools to convert audio data to CLAN transcription files. Because this field of study is quite old, there are many automated resources that could be utilized to create group decision-making technology.
For now, this is just an idea I wanted to write down. If you have any ideas related to this shoot me an email at dcornellpublic[at]gmail.com