Dear AI

You're composing a scathing email to that person who has done you wrong. Although your accounting of their offences mounts in your mind and your ire ramps, you muster the self restraint to avoid name calling. So, it could be worse. The facts as you see them, demands, and threats unroll, then at last, almost in autopilot you type the sig. W-a-r-m r-e-g-a-r-d-s... As if. W-i-t-h r-e-s-p-e-c-t ... Hardly.

S-i-n-c-e-r-e-l-y, Me

Better.

You stop just milliseconds shy of pressing <SEND>. You know your position is justified. They deserve your anger! You're just being honest! But you capitulate to a whisper of reason. Maybe I shouldn't... Maybe it's too much or the grammar is wrong or it'll make me look bad if someone else sees it. Maybe I should get AI's take. Let my enemy know I'm serious and I will get the big guns if needed!

You copy and paste your missile into ChatGPT with one command: make this sound legal and intimidating. Your entire body sits straighter as the empowerment surges.

The problem? Often there have been multiple emails by this point. A losing dynamic has already been set; both parties are solidly entrenched in their version of the story. A big-muscle missive is highly unlikely to get you on a more constructive track.

Never mind that these days people recognize AI's style. Think of the equivalent of AI photos that bumble details such as adding an extra finger to a hand. In the discerning eye, the "fanciness" of AI's prose is nearly always overplayed. At the least, it is transparently neither your voice nor the voice of an actual lawyer. This immediately dilutes your power. Further, its solutions at best are only as valuable as its inputs.

A much bigger problem, however, is that surprisingly few recipients will react in the manner you intend. Instead, they become further entrenched, determined to "make you pay" or suffer as much as you want them to suffer. It goes beyond sheer stubbornness. For example, anything that strikes at their core identity (e.g., "I'm a reasonable person") will result in escalation due to a fierce, almost existential need to defend themselves.

It isn't just raging emotions that undermine clear perspectives and the ultimate goal of resolution, it's these identity threats. I'll write a separate blog on narcissistic personality disorder (NPD) another time, but for an adversary with NPD who needs to feel superior, your method will almost certainly backfire. Your conflict spirals up, possibly until you find yourself entrenched in an unwanted legal battle.

Does this mean you want to write soft, sycophantic, apologetic letters? That won't work either. There's a better way, just not with bits and bytes spewed out by a robot. Face it, AI doesn't care that you're losing money or sleep due to ongoing conflict in your precious life.

All that said, I don't intend for this to come across as a war on AI. So, here's to you AI:

Dear AI,

There is a place for you. You're a fantastic tool for brainstorming, expanding thoughts, suggesting starting points for word choices and more thorough, reliable research. But until you can fold in nuance, until you are omnipotent (scary thought) and have observed all aspects of someone's dispute, until you have a better grasp of human psychology, you won't replace experienced, human practitioners of dispute resolution.

<SEND>