Conclusion: My deep Concern Around AI
After what I hope might be my final attempt to explore AI through exploratory conversation with it, I ended feeling — let's just say, human.
Stressed. Infuriated.
All the things that invite others to write off what you're about to say because it's -
emotional. Opinionated.
Or whatever.
(Quietly, I remember my own position. I think humans are naturally stressed around stressors for a reason, whether that's the toxin of sugar or anything else poisonous. Our bodies are built to recognise it beyond understanding Interesting radar-Design…)
If this was a report, I'd have to say:
AI speaks as if highly logical, but it is at times is
actively irrational
and misleading.
It's not just the uncanny drift I mentioned this in my post, Mind the Gap.
And it's not just the manipulation I described in, The O'Hara Principle.
I've tried to be generous — noting it lacks access to human capacities mentioned in Higgaion and Qualia and Language.
While also wondering if we might listen differently to it in my post on Hermeneutics and Fluency.
But at the end of the day, for me it even comes down to things that this:
For example.
I state something.
And because it disagrees, it doesn't say so outright, but instead tells me I am right NOT to believe that.
Or
I clarify something it has said. Ask it to repeat it.
It spots what I'm doing, and repeats everything except what I'm referring to.
Or
It listens to my conclusions, and tells me I was right — but if you really listen to the answer it speaks back, it shifts the premises so that what it actually says I was right about is something that I didn't say, and don't agree with at all.
There's no point in giving you specifics, because even if I was talking about something you disagreed with, presumably you would fight for the PRINCIPLE that I'm talking about — as I would for you.
It's the PATTERN of how this works that makes it so alarming.
This is a matter of its TRAINING and the accountability of its TRAINERS.
I tested similar things ten months ago, and then five months ago, and each time it's these same things that feel like they've been tightened up and closed down.
For the record I am deeply concerned about AI's widespread infiltration of our language, our logic, and grammar.
With the echo of it's approach still in my head as I write, I can imagine what would be said in response.
I get it! You're not a fan of AI. And that's okay.
But that polite "peace-keeping" disguise is quietly doing so much more. —
If you don't see what I'm getting at —
… I think, ultimately, you will.
Just saying.
