Have you ever found yourself wondering if artificial intelligence could offer an impartial perspective on your personal interactions—be it a heated political debate or a series of late-night text messages with a romantic interest?
In the last week I’ve seen two different intriguing examples via social media about the results of using LLM chats and showcasing AI’s prowess in dissecting human interactions, shedding light on biases and emotional subtleties:
1. Someone put the transcript of the recent Trump vs Harris presidential debate night into AI, and instructed it to analyze it for bias. It concluded the moderators were favorable toward Harris and negative towards Trump and put him on the defensive more than the other candidate. It didn’t take AI to see this for yourself if you watched the debate live.
So this guy put the entire debate transcript in ChatGPT and asked ChatGPT to analyze the moderator bias and every single person needs to hear this because oh my God is it right on the money. pic.twitter.com/aRyythOB6V
— Insurrection Barbie (@DefiyantlyFree) September 12, 2024
2. The other example is of a woman on TikTok who took a screenshot of her text messages with romantic prospects and had AI analyze them to determine certain things in case she was reading things wrong.
This is the article I was reading, Woman Asks ChatGPT To Analyze Her Date’s Texts, Left ‘Shocked’ by Response.
Most of the comments I saw on the Twitter post were mocking the woman in the latter example to being playful at best, but I thought there was something brilliant in the subtext.
You Get Out What You Put In
Obviously AI and ChatGPT-like LLMs can only analyze data that’s submitted to it, like text or images or even the audio to transcribe, and are still as-of-yet severely limited in their abilities to analyze emotion and tone in those things.
From the article,
“The relationship expert knows that while AI-powered tools can offer possibilities for augmenting personal insights, the core of human relationships lies in the nuanced experiences that technology cannot yet replicate.”
However, if we take a “just give me the facts” approach based on only the input, AI chat tools can be very helpful precisely because of how they can’t read emotion and nuance, but only the raw data input, which in this case is written words.
This is precisely how I’ve been using AI a lot in my business and in writing.
In the past year I’ve had multiple instances where clients have tried to scope creep and get me to do more work than I agreed to. I’ve had others assume I’ll do something they realize needs to be done, despite never having discussed it, let alone come to a mutual understanding of my role and the service(s) I had actually agreed to for the compensation.
I do my best to have things in writing (like contracts or agreements) but sometimes both parties and can read the exact same words a different way. Through a different lens.
I do my darnedest to have all my business and client calls over Zoom or Google Meet so I can use the Fathom Notetaker. The free version of this allows you to not just have the call recorded, but store that recording on their site via your account, and provide a free transcript. The cheapest premium version of Fathom lets the user get certain types of analyses like extracting action items from the transcript or use it.
I still use the free version and just copy and paste the transcript into my preferred LLM for this purpose.
Here’s a chat where I asked ChatGPT what are some uses for Fathom Notetaker.
Using AI as an Impartial Judge or Third-Party Mediator
In recent months, to avoid the scope creep problem I’ve had happen a few times, I have taken the transcripts of such business calls and submitted them to AI and have it analyze the call and tell me “what specific terms did both parties agree to, based on this chat”. And sure enough, I get the unbiased factual responses of what was actually discussed as assumptions can’t be extrapolated from the transcript alone. This helps protect me from both myself and the other party from “misremembering” things, but has also allowed me to stand my ground and not be pressured or confronted falsely for something the other party is mistaken about.
As technology continues to advance exponentially at warp speed, AI tools like ChatGPT and Fathom Notetaker are becoming invaluable allies in both professional and personal spheres. While these tools might lack the nuance to fully grasp the things like human emotion and tone of voice, their value lies in their unbiased, fact-based analysis. This can be especially beneficial in business dealings and personal relationships where clarity and precision are paramount.
By leveraging AI to parse through our recorded communications, we can avoid misunderstandings and hold each party accountable to what was actually said ‘on the record’. In a world filled with so much noise, sometimes it’s invaluable to have a tool that can cut straight to the facts.
The tool I’ve been using for the last 18 months for the kinds of things I’m talking about in today’s post is Magai, which allows quite a few of the most popular LLMs and Iterative image generation services in ONE single account. Get a free trial when you subscribe using my link.
In the meantime, since posting this article I came across this video on Instagram which I thought also proved the “using only these parameters, tell me _______” approach:
