Weekly Review 17 November 2023

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

1. Yes Billy, after the crap you directed at Siouxsie Wiles and others, you do deserve to go to jail: https://www.nzherald.co.nz/nz/billy-te-kahika-on-covid-19-protest-prison-sentence-i-shouldnt-have-to-go-to-jail/7MNYLHBQLRCGNBIA5JJCKS3MYI/
2. I moved cities three times for my three post-docs, and each time was hard on my family. Post-docs shouldn't just be for young, single people, but that is the reality: https://www.science.org/content/article/i-didn-t-want-relocate-postdoc-fellowship-reviewers-penalized-me-it
3. Watermarking of images generated by AI can only be effective and trusted if all images are watermarked with their provenance, whether the creator is human or an artificial intelligence: https://www.datanami.com/2023/11/06/can-watermarking-solve-genais-trust-problem/
4. The kind of people who would misuse artificial intelligence to produce this kind of material would certainly not use watermarking: https://www.theregister.com/2023/11/10/child_psychiatrist_sentenced_ai/
5. Another case of misused artificial intelligence where watermarking would certainly not be used by those who created the material: https://www.stuff.co.nz/world/us-canada/301003371/ai-fake-nudes-are-booming-its-ruining-real-teens-lives
6. The people who promulgate conspiracy theories would not be deterred by watermarking of fakes generated by artificial intelligence, or by any other verification technology: https://spectrum.ieee.org/deepfake-2666142928
7. My suburb was cut off by floodwater during the Auckland Anniversary Weekend floods in January 2023. I wonder if this machine learning model could have predicted them? https://spectrum.ieee.org/flood-warning
8. While the author of this article was able to make their way from a deprived start into a research career, I can't help but wonder how many other potentially great researchers have not been able to do the same? https://www.science.org/content/article/homeless-phd-found-way-research
9. Generative artificial intelligence does not yet have the framework of rules and heuristics that we call common sense to back up and refine what it produces. Which gives rise to some odd outputs: https://www.theguardian.com/travel/2023/nov/12/chatgpt-travel-planning-intinerary-walking-ai
10. "If anything can be fake, then nothing has to be real". Has artificial intelligence brought us into a post-reality world? https://www.informationweek.com/machine-learning-ai/deepfakes-get-weaponized-in-the-gaza-war
11. It seems quite obvious to me that the established players in artificial intelligence will push for regulation that favours them: https://www.datasciencecentral.com/regulatory-capture-why-ai-regulation-favours-the-incumbents/
12. This AI pin mostly does what a smartphone does, but adds a wide-angle camera and laser projector. So why not make a clip-on camera and projector that talks to your phone? https://techcrunch.com/2023/11/09/humanes-ai-pin-up-close/
13. Call me a cynic, but could artificial intelligence really mess up the economy worse than Wall Street? There were not AI in 2008: https://www.theguardian.com/technology/2023/nov/09/yuval-noah-harari-artificial-intelligence-ai-cause-financial-crisis
14. Of course, relying on machine learning for stock trading isn't the safest way of avoiding the artificial intelligence economic apocalypse: https://www.datasciencecentral.com/artificial-intelligence-making-waves-in-the-stock-market/
15. Health care is a much better area to apply artificial intelligence and machine learning, compared to finance: https://www.datanami.com/2023/11/10/unlocking-the-potential-of-ai-and-ml-in-healthcare-today/
16. The provenance, and licensing, of data sets used to train artificial intelligence is a bit of a mess: https://spectrum.ieee.org/data-ai
17. OpenAI wants to help build new training data sets for artificial intelligence: https://techcrunch.com/2023/11/09/openai-wants-to-work-with-organizations-to-build-new-ai-training-data-sets/
18. A few years ago, when I was the Head of an IT teaching department, I used to tell the tourism students that artificial intelligence was going to take their jobs before it took the IT students' jobs: https://www.stuff.co.nz/travel/news/301004486/tourism-industry-choosing-robots-over-staff-no-training-no-staff-meals-no-leave
19. A cheat sheet for using ChaptGPT: https://www.techrepublic.com/article/chatgpt-cheat-sheet/ 
20. Adding more parameters to generative AI is not guaranteed to solve the underlying problems facing artificial intelligence, which come from a lack of understanding of the language they are generating: https://www.extremetech.com/computing/report-amazon-is-developing-ai-model-twice-as-complex-as-chatgpt
21. As more organisations adopt artificial intelligence, data centres must adapt to the needs of this technology: https://www.informationweek.com/it-infrastructure/how-data-center-infrastructures-must-change-to-support-ai
22. So if you're good at something already, you don't need artificial intelligence to do it for you? Am I the only one who finds that kind of obvious? https://techcrunch.com/2023/11/10/ai-makes-you-worse-at-what-youre-good-at/
23. Now Samsung is introducing artificial intelligence into its smart phones https://www.theregister.com/2023/11/09/samsung_introduces_generative_ai_on_devices/
24. I suspect the profession of "prompt engineer" is going to disappear very soon, especially as generative artificial intelligence gets better at understanding prompts: https://www.kdnuggets.com/why-prompt-engineering-is-a-fad
25. How will IT departments adapt to the needs of artificial intelligence? https://www.informationweek.com/machine-learning-ai/revamping-it-for-ai-system-support
26. Is robotics soon going to have a "ChaptGPT moment"? https://techcrunch.com/2023/11/10/ai-robotics-gpt-moment-is-near/
27. Detecting infidelity using artificial intelligence: https://dataconomy.com/2023/11/10/cheater-buster-ai-and-how-does-it-work/
28. Generative artificial intelligence brings a whole slew of ethical issues with it: https://www.kdnuggets.com/the-new-ethical-implications-of-generative-artificial-intelligence
29. The lack of transparency in large language model artificial intelligence: https://www.datanami.com/2023/11/09/watchful-plots-transparency-of-black-box-llms/