First Encounters and Thoughts on AI

So I have been seeing the deluge of all things AI in the news and on my feeds etc for a little over a year. This week I was confronted with two AI examples in my own life for the first time.

In text someone whose second language is not even English and had used chatGPT to write a job application letter. He wanted me to look at it and make suggestions and only admitted to me after I had altered it that it was made with AI.

So firstly the english was impeccable, maybe only a few too many commas. But as a letter it was terrible and was so generic as to be immediately ignorable by any employer. I can however imagine with better prompts that a good letter could have been formed.

The second was in an architecture competition which I am doing with a friend. He put some prompts through midjourney concerning some basic ideas we had about the building.

The images were immediately amazing to me on first look. But on deeper inspection they fell apart and were not convincing, strange glitches and impossible artefacts in the building being only the most obvious failures.

I can see these as becoming tools and perhaps shockingly meaning some jobs in which expert knowledge is required will be quite quickly supplanted by AI.

It’s also quite clear from the results that AI doesn’t know what it’s doing (not true AI) but that AI makes accessible brute force knowledge like language knowledge and photorealistic imagery that otherwise for individuals requires thousands of hours to build up their abilities in.

Should I be afraid or reassured? The tools are exciting and potentially for all of us liberating but as usual I never stop being surprised by the depth to which humans can f–k it up.