Additional reading

1. Expert writers on their use of ChatGPT

Guest Opinion Essay published in The New York Times, written by Meghan O'Rourke after her thought experiment with ChatGPT. «When I write, the process is full of risk, error and painstaking self-correction. It arrives somewhere surprising only when I've stayed in uncertainty long enough to find out what I had initially failed to understand. This attention to the world is worth trying to preserve. The act of care that makes meaning – or insight – possible. To do so we will require thought and work.»

2. Homogenization of language

Since the release of The New Yorker article by Kyle Chayka "A.I. Is Homogenizing Our Thoughts" which featured research from colleagues at Cornell University, Santa Clara University, more papers and preprints pointing towards influence of LLMs on human spoken communication are being released. Check the article here.

Check one of the preprints from Max-Planck Institute here: "We apply econometric causal inference techniques to 740,249 hours of human discourse from 360,445 YouTube academic talks and 771,591 conversational podcast episodes across multiple disciplines. We detect a measurable and abrupt increase in the use of words preferentially generated by ChatGPT, such as delve, comprehend, boast, swift, and meticulous, after its release. These findings suggest a scenario where machines, originally trained on human data and subsequently exhibiting their own cultural traits, can, in turn, measurably reshape human culture. This marks the beginning of a closed cultural feedback loop in which cultural traits circulate bidirectionally between humans and machines. Our results motivate further research into the evolution of human-machine culture, and raise concerns over the erosion of linguistic and cultural diversity, and the risks of scalable manipulation."

3. Saved time?

METR ran a randomized controlled trial to see how much AI coding tools speed up experienced open-source developers. Developers thought they were 20% faster with AI tools, but they were actually 19% slower when they had access to AI than when they didn't. For more details, check out their blog post here and paper here.

4. Upskilling or Deskilling?

The use of AI-based imaging in endoscopy results in potential "deskilling" of physicians to identify lesions without AI assistance. Check the paper here.

Distributed under CC BY-NC-SA 4.0