From a legal scholar friend of mine, out West, a few thoughts on Artificial intelligence.
We must adopt the same careful, suspicious, exhaustively specific approach one would use when bargaining with the Fair Folk, Cursed Monkey Paws, or Djinn.
- Ask clarifying questions but don’t trust their answers to be complete
- Test for loopholes and hidden dangers in the bargains you strike with them
- Assume nothing is implied, they are not human and don’t understand the intentions of mortals
- Don’t disable safety parameters or worse yet teach them how to do it
- Most importantly remember that getting exactly what you asked for might be the worst possible outcome
So if the instructions on your AI say not to feed it after midnight and always use a salt circle, do so….you have been warned. Cort Jensen.
Fair note, Cort is a fan of fantasy and science fiction and an avid movie consumer (with brutally honest reviews issued on the regular). Those readers who are not well versed in those area may find that the above as a bit confusing as most find AI. Both the those not familiar with the court of the Fae or the workings of Gronk, Gemini and ChatGPT should endeavor to expand their knowledge base.
The potential for leveraging fire, the wheel, the plow, the cabbed tractor, the GPS and the internet have all been discussed before in prior days. AI is no different. The question remains, Who, what, when, where, and why? In the words of my father’s Vietnam era platoon sergeant to a young draftee one hot day “Think DMNT, even it if hurts”. The lesson applies to using AI. Think about where the data is coming from, who is sponsoring the AI and what do they get out of creating a detailed cartoon of your office staff to put on Facebook?
