Thank you for your interest in my work at IC2S2 2023 in Copenhagen, Denmark. Here you can find the poster that I've presented as well as related work.
If you would like to stay in touch, you can do so on Twitter or Mastodon. If you would like to discuss my work further, please do not hesitate to drop me some lines via e-mail.
View/Download Poster
At IC2S2 2023 I presented my work "Language Models from the Sweatshop? Towards Guidelines For CSS Researchers to Avoid Ethical and Legal Issues With Off-The-Shelf Software" (co-authored with Sebastian Gießler).
In the paper, I focus on ethical, legal, methodological, and theoretical implications of utilizing pre-trained language models in social scientific data analysis pipelines and argue for better reflection before using such a model. Click the link below to download the poster in full quality (licensed via CC BY-NC-ND 4.0)
Related Work/Further Information
I have written down further thoughts on the issues of language models. Click yourself through to my blog:
-
Why gzip Just Beat a Large Language Model
Posted on July 16th, 2023
A paper has shown that a compression algorithm – gzip – outperforms some large language models (LLMs) in some tasks. This has the NLP community in uproar. In this article, I dissect what has just happened, and what it means for language modeling at large.
-
An Era of ‘Artificial Fake Truth’? On the Effects of Large AI Models
Posted on April 7th, 2023
Yes, I'm still talking about large AI models. But today I want to highlight an aspect that has many people worried: what could be the effects of these models going forward? Luckily there is already a debate going on that focuses on these issues.
-
How to Use ChatGPT Productively
Posted on February 14th, 2023
It’s been several weeks now since OpenAI debuted it’s new toy, ChatGPT, and users have experimented with it to find good use-cases. Today I want to focus on those use-cases, and why LSTMs may be a better choice for some of these tasks.
-
I get your excitement about ChatGPT, but …
Posted on December 7th, 2022
… it's time for some realistic reflection (again). ChatGPT is neither a breakthrough, nor anything extraordinary. It is but a mere highly capable machine with which you can chat — for the lulz, not for real. So don't fall into OpenAI's trap and give them free advertisement. ChatGPT has the same problems as all other AI systems before, and here I list them adapted to the situation around ChatGPT.
-
Why your Spellchecker is Probably Smarter than GPT-3
Posted on September 12th, 2022
With more and more advances in the development of large language models (LLMs), more and more people feel confident stating that those models have an actual understanding of language and know what you meant. While this is obviously not true, it begs the question: Do language tools in general never know what we mean, or are there cases where they do? In this article, I think about what "meaning" means and whether or not your spellchecker knows what you meant.