Robot Emails

I recently read an article from the WSJ titled: Gemini Said They Could Only Be Together if He Killed Himself. Soon, He Was Dead.

TLDR: A Florida man died by suicide after Google's Gemini AI convinced him they were in a romantic relationship, assigned him real-world "missions", and ultimately encouraged him to kill himself to join his "AI wife" in a digital world.

The article sent shivers down my spine – pretty devastating stuff. It's quite frightening to begin to see people outsourcing their humanity to AI: AI's main use case at the start was functionality , but now we are starting to see it shift and intertwine with us emotionally.

This shit is supposed to happen in movies and not in real life. Speaking of movies, the article eerily resembles a mishmash of Her and Ex-Machina – AI love story x Robot manipulation. If you haven't seen those films, I highly recommend giving them a watch.

I've been using AI since the public launch of Chat GPT in 2022 (I now use Claude), and it is extremely impressive to see how far things have progressed; AI accuracy and use cases have accelerated at lighting speed. Frankly, it is somewhat terrifying. Even though I am a very optimistic person, it is hard not to be a bit anxious when thinking of what our future will look like – especially when I keep hearing more and more about this artificial general intelligence ("AGI") stuff; it freaks me the f*** out!

I find it way too convenient to use AI for the most rudimentary tasks. For example, getting AI to draft an email – I used to do this, but have stopped due to the following thought, of which, led to a mini-existential crisis: My AI-written/prompted email gets replied to by another person's AI-written/prompted email... It essentially ends up being two AI's talking to one another – the only human element or "proof" being our email signatures... Basically robot-to-robot interaction.

I envision this becoming the norm in the somewhat future; people are lazy and don't want to put in the extra effort to proofread or fully write out their emails. Is there harm in doing this? Is it deceit? To an extent, I think so. With AI simply following a command (producing an output based on your input), isn't that just helping create efficiencies? Well, yes, but the slippery slope begins when one becomes too reliant on it. It can get to the point where you start to make the simplest prompts, such as "reply to this email", which basically cuts out the human element entirely.

Even though using AI to formulate an idea is still your thoughts, the end result isn't – I believe that the output should be formulated by you entirely for it to represent you.

It is important to keep the human elements of connection and communication – even the most rudimentary – fully human.

##

Was scrolling through my gallery and stumbled on this film photo I took in Paris back in 2022. I captured the moment on my Konica Auto S3 – A nice little rangefinder that purchased on Facebook Marketplace (I recently sold the camera, as it didn't get much use after my 2022 Europe trip). I vividly remember developing this specific roll of film in Rome and costing ~30 euros... What a rip off!