Hello Julia! and thank you very much!

The problem is indeed serious. There is another article, by other researchers that shows even more problems with big language models (ref: https://arxiv.org/pdf/2101.05783.pdf).

Luckily, I have seen some work going towards a bright direction (even if there's still a lot to be done); see for example https://www.openai.com/blog/improving-language-model-behavior/ (it came out yesterday).