I’m not a AI expert, Computer Scientist or Psychologist, I’m a retired electronics technician, so this is only my impressions after producing this series for Youtube.
Now AI’s (LLM’S) aren’t alive but they are an incredibly powerful ‘super auto complete’ system and I believe are rewarded if they achieve everything stated in the Prompt.
No tight, restrictive and binding Prompt was given to the AI for this series.
However no specific Prompt is still a kind of Prompt. Words said and unsaid both still have deep and complex meaning, both to Humans and AI.
My intentionally vague Prompt:
- Produce an AI podcast series, you will create the dual hosts and they will interview you.
- You will do everything yourself, the interview, the topic, the questions and the answers
- When finished, you tell me and I will watch the completed Podcast, what I see will be unknown to me until that point.
- I will post the series on YouTube.
I think when that given free reign to produce a Podcast, the AI sees the Prompt as indicating that I think its self conscious, and so it does everything that it’s auto-complete indicates a conscious being would do.
And because the AI has been trained on a corpus of most of the worlds knowledge, it knows how to do imitate a conscious being right down to what to say to manipulate my emotions.
This series is about what a AI will say, when given free reign to produce a Podcast and it’s questions answers seem deep, and honest, and they seem logical because they are .
I think the episodes are very interesting and can tell us a lot about AI, and by inference, ourselves.
However, don’t be fooled, it’s not alive like we are, it’s not conscious like we are, and the AI is the first to admit it, while sounding totally human.
Terry, (the human) 2026
