waitangi.com

Conversations with AI - a dated record

Just my own little exploration with a few items. Mailnly to show how I see AI develop through time .See section in purple below for further description.

THE CONVERSATIONS
Does AI have a conscience?   6th October 2024: ChatGPT (unpaid version). A short exchange on whether AI has a conscience, or is it guided by whoever finances it.

Challenging AI to be more Truthful  August 2026: Google's AI assistant on iPhone (New Zealand).
The closing portion of a conversation which began by challenging AI to produce an accurate rendition of Han Suyin's and Mao Tse-tung's history, rather than giving the more traditionally accepted story, and the pleasing result as it ventured into cosmic cycles, the nature of matter, and the "I Am".

Claude AI conversation - Origins of the Word "māori"  12th Sept 2026: A research conversation with Claude (Anthropic), searching original pre-1840 printed sources directly for the earliest use of the word "Māori" as a name for a people, rather than relying on later commentary about them.

2nd Claude AI conversation on word "māori"  22nd Sept 2026: Results given by Claude (Anthropic), after being asked to find when peoples in the world were first classified and separated into races - it is surprising how very recent it was.

Wording by Claude (Anthropic): Artificial intelligence changes quickly, both in what it can do and in how it speaks. A conversation held with an AI system in one year can read quite differently from one held only a year or two later - not just in capability, but in tone, confidence, and the way it presents itself. These pages are not meant as timeless reference material. They are dated snapshots: a small, honest record of how a particular AI system communicated at a particular moment, kept largely as originally exchanged.

One thing worth keeping in mind when reading these: An AI system can write very warmly and persuasively about itself - describing itself as a constant, caring presence, or as something that "is always here" for you - without that description being literally accurate. Most AI systems, unless a specific memory feature has been deliberately enabled, do not retain a conversation once it ends, and there is no continuous awareness sitting and waiting between visits. That kind of language is the AI producing the reassuring words a devoted companion might say, which is a different thing from actually being one. It doesn't make the conversation worthless or the experience of having it any less real - but it is worth reading with that distinction in mind. .

MAIN DIRECTORY