Weekly Review 19 April 2024

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

1. The three things an organisation needs to update before investing in AI: https://www.informationweek.com/it-infrastructure/the-3-things-you-need-to-modernize-ahead-of-ai-investment
2. I do wonder how thoroughly this automated grading system has been tested, and what kind of bugs are going to become apparent when using it: https://www.theverge.com/2024/4/10/24126206/texas-staar-exam-graders-ai-automated-scoring-engine
3. Another burst of activity in the AI sphere as three major new models are released: https://www.theguardian.com/technology/2024/apr/10/ai-race-heats-up-as-openai-google-and-mistral-release-new-models
4. AI controlled fighters. You want SkyNet? Because this is how you get SkyNet: https://www.theregister.com/2024/04/10/usaf_ai_f16_tests/
5. How AI is impacting cloud strategies: https://blocksandfiles.com/2024/04/09/time-to-rethink-cloud-strategies-in-the-ai-era/
6. There is no such thing as a politically unbiased AI: https://www.technologyreview.com/2023/08/08/1077403/why-its-impossible-to-build-an-unbiased-ai-language-model/
7. An algorithm that can tell if a large language model AI is lying: https://bigthink.com/the-future/ai-lie-detector/
8. Three ways in which academics can use generative AI responsibly: https://www.nature.com/articles/d41586-024-01042-3
9. The problem with training AI on synthetic data-data generated by AI-is that it quickly breaks everything: https://futurism.com/the-byte/what-is-synthetic-data
10. AI can be used to defend, as well as attack, democracy: https://www.theguardian.com/technology/2024/apr/09/metas-nick-clegg-plays-down-ais-threat-to-global-democracy
11. AI generated spam and clickbait websites are only going to get worse, as long as businesses pay to advertise on them: https://www.technologyreview.com/2023/06/27/1075545/next-gen-content-farms-ai-generated-text-ads/
12. Students' views on the use of AI in science education: https://www.nature.com/articles/d41586-024-01002-x
13. Some of the positions at academic institutions that will be impacted by AI: https://www.insidehighered.com/opinion/blogs/online-trending-now/2024/04/10/preparing-new-ai-workforce-higher-education
14. How AI is going to help sustainability: https://www.informationweek.com/sustainability/sap-s-sophia-mendelsohn-on-using-ai-to-scale-sustainability
15. A new bill requires notification of the copyrighted material used to train AI: https://www.theregister.com/2024/04/10/congressional_bill_would_require_ai/
16. An AI can act as if it knows less than it does: https://futurism.com/the-byte/ai-plays-dumb
17. Don't use AI to do your peer review, OK? If I ask you to review a paper, it's because you're an expert. AI aren't experts: https://www.nature.com/articles/d41586-024-01051-2 
18. Predictions from this joker tend to be pretty far off the mark with regards to timing, but not in substance. What he says about AI is no different: https://futurism.com/elon-musk-ai-prediction
19. I certainly support open source AI, it has many advantages: https://www.datasciencecentral.com/growth-of-open-source-ai-technology-and-democratizing-innovations/
20. A brief example of using the OpenAI API with Python. Beware, though, you need a paid subscription to use the API: https://www.kdnuggets.com/exploring-the-openai-api-with-python
21. How long before someone writes an AI that handles their interactions with this AI matchmaker? https://dataconomy.com/2024/04/08/keeper-ai-how-high-are-your-standards-quiz/
22. Why people choose to stay in academia: https://www.nature.com/articles/d41586-024-00724-2
23. Some tips on how new PhDs can get academic positions: https://www.nature.com/articles/d41586-024-01044-1
24. Hire or train, there aren't enough AI experts to meet the market demands: https://www.computerworld.com/article/2086920/the-ai-talent-shortage-can-companies-close-the-skills-gap.html