Some problems with AI
Here are some problems that I think hound our use of AI:
- Our ability to discern what is good or terrible. With all the slop and seemingly perfect prose we now increasingly and perhaps unwillingly consume, we may lose our “taste” for good writing.
- Our ability and willingness to verify. If we ask AI to analyze a huge dataset to uncover insights and make discoveries, can we easily check the accuracy of the output? It may be easy with less complex tasks, but how about more complex ones that will burden our brains?
These problems are more pronounced in the unthinking and the lazy.
Some practical prescriptions for now:
- Continue to hone and show “good taste”. How? Read more excellent, human-produced writing, both fiction and non-fiction. I know it’s hard work, but I don’t see how you can develop good taste with AI-produced dross. Seek genuine human experience.
- Raise an eyebrow and validate. A healthy dose of skepticism is a necessary attribute of a functional human being. This is not only essential for complex analysis. For more mundane tasks, subtle mistakes and omissions can compound. They can also be embarrassing.
Do these make sense? What problems do you see with how people use AI?