AI and Conflict Mitigation

Recently, I met Ed Gomes, my long-lost college classmate, at his Montana ranch, where he lives a socially isolated life. I suggested he join our class WhatsApp group. Here is the conversation that ensued:

Ed: Before I join, may I browse the group chat on your phone?

Me: Sure. Enjoy!

Ed: I read the chat messages from three months prior. This group is consistently predictable and plagued by conflict. Don’t tell me you like the group!

Me: I love it. There’s comfort in predictability, and you can’t have conflict-free groups. Unfortunately, discussions of certain topics and personality clashes sometimes create dysfunctional conflict. Consequently, some members leave the group in disgust.

Ed: What has the group done to manage conflict?

Me: After some unsuccessful band-aid fixes, we are onto something promising.

Our class valedictorian, Neel Gupta, who has an artificial intelligence (AI) startup, is pilot testing his AI system on our group. For this system, he obtained the chat archives, detailed dossiers on members, and their responses to a survey designed to understand the psychology and sociology of their lives.

After the AI system kicked in, it blocked us from posting in the group chat.

Ed: But the chat is going on!

Me: For the past three weeks, the chat has been AI-driven. The AI system generates and auto-populates messages on behalf of group members and their interactions regarding politics, sports, movies, forwards, special occasions, and everything else. It feels like before, except that everything is AI-driven.

Ed: How’s that even possible?

Me: AI has come a long way. It drives our cars, guesses our internet search queries, and auto-populates our chat groups. Of the many messages you browsed, some were AI-generated, and you didn’t notice any difference. Group members have had no issues with the content and tone of messages posted on their behalf.

Ed: What about new personal stuff that the system cannot know?

Me: There’s a link where we can input information about any personal events/activities we wish to share. For instance, I could input I visited Spain and attach any pictures I wanted to share. The system will determine how and when to present it in the chat.

There’s a plug for every concern you may have about the system.

Ed: How does it feel to witness the drama of your life playing out in the group chat?

Me: It’s surreal—seeing life as an illusion, puppeteered by a supernatural-like entity. This setup brings to mind philosophical ideas like dualism and Lila.

Ed: Philosophical notions aside, how will the AI system function as a conflict mitigation tool?

Me: The system will contribute to ebbing hostility in the minds of group members in three ways.

First, the negative message source effect, which is very strong in our group, will weaken considerably.

The source of a message affects how we perceive a message. For example, identical words of advice from my mother and mother-in-law evoke different responses from me. Likewise, my response will differ when I read something posted by someone, who I think is obnoxious, versus reading the identical message posted on her behalf by the AI system. I might find the latter less bothersome.

Second, the AI-driven chat will be an eye-opener. We typically see ourselves through our eyes. We are very reluctant to look at ourselves through the eyes of others, but sometimes, we should.

The AI system makes us see ourselves through the eyes of others, in this case, a technological system. This perspective could be beneficial for some. For example, consider members who routinely and loudly denounce hate and divisiveness but ironically whose forwards/messages are driven and blinded by hatred. They would benefit from seeing themselves through a different set of eyes.

Third, this unique experience could be thought-provoking and eventually transformative for some. Perhaps, it will push us to contemplate the big picture in life more than before. Have a sobering influence on us. Make us less biased and less rigid. All of which will lead to fewer disagreements and less conflict.

Ed: But, conflict, as reflected in the content and tone of chat messages, would become inconsistent with the gradual reduction of hostility in the minds of group members.

Me: The AI system is dynamic. It will align the content and tone of the chat to system-predicted changes in the mindsets of group members. The system will iterate toward an optimal state of healthy functional conflict.

Ed: This stuff is mind-boggling. And scary.

Me: There is nothing scary. A brave new world awaits us. Just embrace it.