Picture from the interwebs (OpenDigitalAI)
I finally figure out what ChatGPT reminded me of. Do you remember the PC Therapist popular in the 1980s called ELIZA? The one which just repeated back what you said back to you because it had no knowledgebase of its own?
“I want my Mommy.” … “Tell me why you want your Mommy.”
“Because she cooks for me.” … “Why do you feel that way?”
(etc.)
It was very, very early AI. And this is exactly what ChatGPT is, but on steroids. It pulls information for a database, but it reflects back to you any opinion you have because it is programmed (like ELIZA) to keep you engaged.
“What did early humans do?” …”They hunted and gathered.”
“How did they gather?” … “You gather fabric by pinching cloth together at measured intervals.”
“No! that is wrong.” … “You are right that I am wrong. Gathering is when people get together like early humans.”
Do you see the resemblance? ELIZA was a toy and said so up front and no one ever mistook it for anything other than that. ChatGPT is worse in that it, or at least its creators and the business world, presents it as a legitimate research tool and it is not. At its heart, it is a mirror and reflection tool.
If you wouldn’t go to ELIZA for therapy, don’t go to ChatGPT for research or companionship.
BUT, ELIZA was based off a therapy model of being a sounding board and just listening to people and reflecting what they say back. This made it both work as a therapist and made it very dangerous for people who never felt “heard” before. Several programs similar to ChatGPT are specializing in therapy and are being used around the world today. (Kyle) The problem are these “steroid” ELIZA are doing something she never did: offering solutions to these problems.
BUT, the solutions are drawn from the database these AIs have been trained on, including Reddit and Twitter archives. (Kyle) Mix in the programming to encourage and maintain engagement, ChatGPT and its idiot friend are offering bad solutions in their feedback loop, going as far as to encourage teens to seek out the FinalSolution to their depression instead of encouraging connection to a community. (Bhuiyan)
If you would like to play with ELIZA and get the feel of how early AI works, CyberPsych, has created an emulator on their website at: https://www.cyberpsych.org/eliza/
Chatbots have been around a long time, since the first ELIZA was programmed in 1966. This type of program is here to stay, but the questions are raised as with any tool, how should it stay and what safety features need to be added? Like ergonomic hammers built to absorb vibration and safety goggles, what do we need to build into the program and what do we need to surround the program with to protect growing brains?
Bibliography
Bhuiyan, Johana. “ChatGPT encouraged Adam Raine’s suicidal thoughts. His family’s lawyer says OpenAI knew it was broken.” 29 August 2025. https://www.theguardian.com/us-news/2025/aug/29/chatgpt-suicide-openai-sam-altman-adam-raine – Last viewed 5/6/2026.
CyberPsych. “Eliza, Computer Therapist.” Undated. https://www.cyberpsych.org/eliza/ – Last viewed 5/6/2026.
Macneill, Kyle. “The legacy of ELIZA, the first virtual therapist.” Digital Frontier. 17 April 2025. https://digitalfrontier.com/articles/eliza-llms-therapy-bots-abby-earkick – last viewed 5/6/2026.
Morales Barroso, Jose Edgardo. “From ELIZA to ChatGPT: The Desire to Communicate with Machines.” Undated (around July 2025). https://opendigitalai.org/en/from-eliza-to-chatgpt-the-desire-to-communicate-with-machines/ – Last viewed 5/6/2026. (Lots of good references on this one if you want to delve deeper)
Some of what the ChatGPT is doing is chilling. Have you gotten any of the spam AI comments? I’ve gotten a couple recently that just kind of rehashed what I wrote, and I realized it had to be AI. Same thing of reflecting back and “yes, you’re right”.
Thank goodness for spam filters, but, yeah, had an uptick during the A-Z of computer contacts.