Smart (part 5) - the not-so-secret AI agents

First an apology: I know I said I would finish this series with this post, but the word agent started coming up everywhere. I had to explore (just a little more) the latest ways we are being convinced (aka tricked) that a sophisticated program is an agent with human-like intelligence.

In part 4, I explained how the first required step in convincing someone that a thing is intelligent is to convince them that it is an agent – an individual with potential agency over its own actions, behaviours and responses. I highlighted what I called ‘human hacking’: the covert and deliberate programming of Smart ChatBots to use language that implies the existence of an agent, and then relying on humans to do what they always do: impute agency, intelligence, sentience, and even consciousness.

Well, now the techies have abandoned the covert methods, to openly declare that 2025 is going to be the year of the AI agent. 

And here I was thinking 2025 was the Year of Glaciers' Preservation; Peace and Trust; Quantum Science and Technology, and Cooperatives.