Action-driven LLMs Will look a lot like #AGI jmcdonnell.substack.com/p/the-near-fut…

Yes, ChatGPT is amazing and impressive. No, OpenAI has not come close to addressing the problem of bias. Filters appear to be bypassed with simple tricks, and superficially masked. And what is lurking inside is egregious. Abeba Birhane Sam Altman tw racism, sexism.

ChatGPT is just mass scale user testing in final stages of GPT-4 prep right?

#Designing #AI to support community wellbeing: our new publication discusses a case-study of a complex system that supported #wellbeing during the #COVID19 pandemic, and how it could serve as a framework for future AI systems. Learn more at frontiersin.org/articles/10.33…

If platforms shouldn't optimize just for engagement, what should they optimize for? We designed an experiment to test selecting Feed content for long term user value, a collaboration between an international group of researchers and Meta. 1/3 wired.com/story/platform…

Marius Hoggenmüller 🏳️‍🌈 presenting our recent work on "Creative AI for HRI Design Explorations" coauthored also with Willem and Kazjon Grace. marius.hoggenmueller.com/creative-ai/ #hri2023

This is very promising news for the #academic research community aisnakeoil.substack.com/p/openais-poli… #LLM #OpenAI

Just arrived in #Hamburg for #CHI2023

🌐🔓#chi2023 opening keynote: remember to change your passwords folks!

Our preprint on the key challenges for #designing #PositiveAI just went online! Find me at #chi2023 if you’re interested #wellbeing #positivedesign #ai arxiv.org/abs/2304.12241

Interested in Generative AI & design research? Submit to our @acmdis2023 workshop July 11. Explore ethical considerations, best practices & future directions. Help create roadmap for integrating into design research. tinyurl.com/genaidis2023 #GenAI #designresearch

Have you been thinking on how Generative AI will impact design research and practice? Come discuss w us at our @acmdis2023 workshop -> Ethical considerations, best practices and future research directions ACM SIGCHI studiolab.ide.tudelft.nl/studiolab/gena… Evert van Beek (@[email protected]) Willem

Following the statement on AI extinction risks, many have called for further discussion of the challenges posed by AI and ideas on how to mitigate risk. Our new paper provides a detailed overview of catastrophic AI risks. Read it here: arxiv.org/abs/2306.12001 (🧵 below)

New preprint on "Positive AI: Key Challenges for Designing Wellbeing-Aligned Artificial Intelligence" - we identify 12 key challenges to designing human-centerd #AI & propose a research agenda to advance scientific understanding of AI's wellbeing impact. arxiv.org/abs/2304.12241

Our own Willem presenting his work on wellbeing and AI alignment (because , what else?!) at Design & AI at TU Delft

How can we orient AI systems towards peace? We don't want them to escalate conflicts, either as a side effect of optimization, or because they can be used by bad actors. Here's a three part strategy, from my talk at Beyond Moderation. docs.google.com/presentation/d… 1/n

New YouTube video: 1hr general-audience introduction to Large Language Models youtube.com/watch?v=zjkBMF… Based on a 30min talk I gave recently; It tries to be non-technical intro, covers mental models for LLM inference, training, finetuning, the emerging LLM OS and LLM Security.

It was really nice to be back in Rotterdam and to give a talk at ThingsCon on AI unintended consequences. An inspiring day full of projects and activities! Looking forward to 10th year edition!

Do you wish you could test a different social media algorithm? Now you can. Announcing the Prosocial Ranking Challenge. $60,000 in prize money -- and we will test five winning algorithms on real users and real platforms using a browser extension. humancompatible.ai/news/2024/01/1…

Valentine's Day could make or break your relationship. Don’t show up with some wack ass flowers or chocolates. Here are 8 reasons you should get your special someone a Vision Pro instead: 1) It'll block out the hoes