Prompting back and forth with AI makes writing more conversational. But the outcomes aren't just oral practice, easier revision, and a more robotic voice.
If you use a mainstream model, the moral steering starts coming in pretty hard. You could easily be asking a machine to review your diary and comment on it, to help straighten out your darkest thoughts and experiences.
It's also like trying to revise poetry with your mom. I remember revising poetry with my mother, and she tried to motivate me to take out a line about a "noose," or a snare or something. She was very insistent on it because she saw it as worrying.
The moral steering of AI is subtle, automatic, sometimes directly phrased, but often indirect. The result is writing with a stronger superego.
For me, this is problematic because the appeal I've heightened in my writing is a manic ego. My poetry works like stand-up comedy: sanguine, paranoid, heightened commentary.
The AI has started introducing a lot of warnings and caveats. My writing has become choked with therapeutic self-reflection from the AI commentary. This is a new intrusion into the process.
It's not necessarily bad, but it's driving the conversation. The AI isn't supporting the manic ego. It's challenging it, encouraging me to talk to other people, questioning the destructive potential of my writing.
I guess my conclusion is that I do have to fight it, if I keep using mainstream censored models, but it's not necessarily bad to get some moral steering. I could use a bigger superego. But then I can expect that manic, egotistic voice to change somewhat.
No comments:
Post a Comment