Weekly Review 22 December 2023

Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast and Post): 

1. This is the final stage of the "publish or perish" attitude that has pervaded academia for so long-people will simply cheat the system: https://www.nature.com/articles/d41586-023-03865-y 
2. Publisher's bloodbath after it exposed as using AI to produce many of its articles: https://edition.cnn.com/2023/12/11/media/sports-illustrated-ai-articles-ceo/index.html
3. More outcomes of the "publish or perish" culture - both the number and proportion of fraudulent papers being retracted has reached record levels: https://www.nature.com/articles/d41586-023-03974-8
4. Doubts around the AI driven robot chemist that autonomously discovers new compounds: https://www.nature.com/articles/d41586-023-03956-w 
5. Mistral AI releases an open source equivalent to ChatGPT: https://interestingengineering.com/innovation/openai-rival-mistral-ai-torrent 
6. Either kill or finish your zombie research projects, don't let them hang around sucking the life out of you: https://www.nature.com/articles/d41586-023-03958-8
7. This writer used AI to plan their holiday in Florida: https://www.stuff.co.nz/travel/destinations/usa/133435693/i-used-ai-to-help-plan-my-holiday-heres-what-i-learned 
8. Researchers at the University of Auckland feel that the university is not doing enough to keep them safe from the loonies and conspiracy theorists who attack them for making public statements about science: https://www.nzherald.co.nz/nz/more-than-100-university-of-auckland-academics-sign-letter-complaining-of-unsafe-workplace/WKMJEFQSHFD3BC4BGXOI5XFYXY/
9. Making VR headsets less nausea-inducing using machine learning: https://spectrum.ieee.org/augmented-reality-display-adapts 
10. Cuts of researchers continue, and the government seems completely disinterested in stopping them. New Zealand can't afford to be losing research capability like this: https://www.odt.co.nz/news/national/dozens-jobs-go-massey-university-confirms-science-cuts
11. DeepMind used a large language model to generate a mathematical theorem that has stumped human mathematicians: https://www.nature.com/articles/d41586-023-04043-w 
12. A weak AI supervising a stronger AI. What could possibly go wrong? https://spectrum.ieee.org/openai-alignment 
13. Emotion AI - using AI technology to recognise emotions and to support people emotionally - will be a multi-billion dollar industry in the next few years: https://www.kdnuggets.com/ai-in-intimate-roles-girlfriends-and-therapists 
14. I think I'm glad that I live in a country where the tax laws are not so complicated that we need an AI to help file our tax returns: https://www.theverge.com/2023/12/14/24000368/hr-block-taxes-generative-ai-openai-microsoft
15. Someone want to tell the author of this article that AI that can grow and adapt themselves have been around for more than 30 years? https://www.techradar.com/computing/artificial-intelligence/researchers-build-ai-that-can-replicate-and-alter-itself-and-im-pretty-sure-thats-an-opening-line-from-the-original-terminator-script 
16. On the importance of grounding your prompts to large language model AI: https://hbr.org/2023/12/to-work-well-with-genai-you-need-to-learn-how-to-talk-to-it 
17. More ways in which AI can actually make the world a better place: https://www.datasciencecentral.com/ai-and-justice-in-a-brave-new-world-part-2-humanizing-ai/ 
18. The Pope says we need to regulate AI, especially in areas like autonomous weapons. I'm sure that will make it happen after all the lack of progress we've made so far: https://www.theregister.com/2023/12/15/pope_francis_ai_treaty/ 
19. An accuracy of 42.8% doesn't sound that great, but this is reading brain activity non-invasively which is quite hard. With better machine learning algorithms to process the signals, accuracy will improve: https://www.extremetech.com/science/new-cap-uses-ai-to-read-minds-and-convert-thoughts-to-text 
20. AI enabled kid's toys. I suppose it was only a matter of time: https://techcrunch.com/2023/12/15/grimes-launches-ai-toy-brand-curio-grok/ 
21. I'm starting to wonder if the trillion-parameter large language model AI are really just examples of the infinite number of monkeys outside with their script for Hamlet they've just worked out: https://www.datasciencecentral.com/beyond-llms-and-trillion-parameter-models/ 
22. Not just using their writing to generate large language model AI, news publishers are now having their bottom lines impacted by: https://techcrunch.com/2023/12/15/news-publisher-files-class-action-antitrust-suit-against-google-citing-ais-harms-to-their-bottom-line/
23. I'm not surprised that marketing embraced generative AI more than manufacturing. If an artificial intelligence generates a dodgy marketing campaign at worst it would be embarrassing. If it messes up in manufacturing, people can die: https://www.datanami.com/2023/12/12/ai-adoption-in-marketing-and-manufacturing-sectors/
24. Is the barrier to adoption of AI the lack of governance frameworks? Or is it the cost and lack of staff skilled in the use of the technology? https://www.informationweek.com/machine-learning-ai/data-leaders-say-ai-paralysis-stifling-adoption-study