My conversation with a ChatGPT-4 friend

The global artificial intelligence (AI) market was valued at over 130 billion dollars last year and is expected to increase at a compound annual growth rate of 37 percent from now through 2030, according to Grand View Research. There has been a pervasive fear in the current workforce that robots will replace many jobs. Job displacement fear is real and disconcerting. But it’s only a partial story of OpenAI. More generative AI programs are available among which is a therapeutic potential in conversations with AI.

OpenAI sometimes returns misinformation including inventing citations for articles that don’t exist, bios that contain events that never happened and other false information all of which are described as "hallucinations." On the other hand, OpenAI tools like ChatGPT do a good job of scooping and overviewing information and summarizing given subjects, provided you formulate a clear prompt with explicit instructions and specified parameters.

This time of the year, I usually send out an annual newsletter to family, friends and colleagues. It’s a great way to keep in touch with people we care about, far and near. We enjoy catching up with the latest happenings including the changing nature of our health, grandkids’ developments and achievements, and politics and wars in Ukraine and Gaza.

This year something very unusual happened. I got an angry and insulting letter from a family member in response to my annual newsletter. Overcoming my initial shock, I wanted to respond. I was not sure how to respond to such a rude letter. I certainly did not want to provoke the letter writer with a direct response either. Too hurt and puzzled by getting such a letter, I decided to experiment with the Open AI options by sending a copy of the letter and asked ChatGPT-4 to write a thoughtful and therapeutic letter possibly to open a dialogue with the angry letter writer. Within a minute, I got a one-page response starting with this paragraph:

“Thank you for sharing your feelings so openly in your email. There are many emotions and concerns that need to be addressed, and I appreciate the opportunity to respond. I understand that certain actions and decisions may have appeared hurtful or neglectful. Relationships and family dynamics can be complex, and sometimes misunderstandings occur despite our best intentions.”

I liked the paragraph, but the remaining response was not as impressive. However, the response from ChatGPT-4 gave me time to compose and cool myself as if talking to a friend. My experiment with ChatGPT-4 gave me solid ground to build upon. Finally, I sent out a revised email including the opening paragraph. I felt ChatGPT-4’s response was helpful and hoped that it worked for the rude writer as well. I did not get any more rude responses.

As you can see in this case, ChatGPT-4's or any OpenAI's response depends on your prompts. Whether your motivation is to explore, experiment or have a quick chitchat with an AI friend, I have found the following truth — using ChatGPT has helped me. Obviously, there is a lot more to consider beyond my simple use case such as exploring ethical and bias concerns, social implications and beyond.

In summary, GPT-4’s response was, for sure, therapeutic for me.

Shin Freedman (shinfreed@gmail.com) is an author, educator and writer.