Weekly Review 12 January 2024

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

1. Using machine learning to improve hardware testing: https://spectrum.ieee.org/hardware-hacking
2. I suspect that these copyright cases are going to come down to whether the courts consider the resulting model to be transformative: https://www.informationweek.com/machine-learning-ai/what-the-nyt-case-against-openai-microsoft-could-mean-for-ai-and-its-users 
3. Yet another copyright suit from authors whose work was used to train AI: https://www.theverge.com/2024/1/5/24027462/another-one
4. Where regulation of AI is heading this year: https://www.technologyreview.com/2024/01/05/1086203/whats-next-ai-regulation-2024/ 
5. How lawyers expect AI will affect their profession: https://hbr.org/sponsored/2023/12/a-new-survey-reveals-how-legal-professionals-expect-ai-to-impact-their-work 
6. It looks like this year is going to be the year of copyright claims against AI companies: https://spectrum.ieee.org/chatgpt-new-york-times
7. This article claims that New Zealand's values of fairness means that we could lead the world in the use of AI. To be honest, 30+ years of neo-liberal economic policies have largely eliminated the fairness in New Zealand: https://www.thepost.co.nz/a/nz-news/350140670/big-opportunity-aotearoa-should-seize-around-ai
8. Sports analogies describing the different kinds of AI: https://hbr.org/2023/12/a-sports-analogy-for-understanding-different-ways-to-use-ai
9. How machine learning was used to diagnose illnesses from the sound of coughs: https://www.technologyreview.com/2024/01/05/1086196/ai-powered-apps-are-tracking-the-sounds-of-sickness/
10. It is a fundamental principle of AI that biased data will lead to biased models. Some approaches for improving data quality: https://www.informationweek.com/data-management/feasting-on-high-quality-ai-data
11. There is not as yet such a thing as a secure AI: https://www.theregister.com/2024/01/05/nist_ai_security/
12. Improving computational performance of AI by combining storage and computation in the same chips: https://blocksandfiles.com/2024/01/04/storagex-wants-to-end-storage-and-compute-separation/
13. An artist quits over the use of AI generated artwork - do they use it or not? https://www.eurogamer.net/magic-the-gathering-artist-quits-you-cant-say-youre-against-ai-then-blatantly-use-it 
14. Guardrails. AI really need guardrails. https://spectrum.ieee.org/in-2016-microsofts-racist-chatbot-revealed-the-dangers-of-online-conversation
15. So adding AI to glasses makes wearables sexy again? https://www.computerworld.com/article/3711844/2024-will-be-the-year-of-ai-glasses.html 
16. What the AI copyright lawsuits are really about: https://www.theregister.com/2024/01/03/kettle_ai_copyright/?td=rt-3a 
17. The different ways in which AI can generate images: https://dataconomy.com/2024/01/08/how-does-ai-generate-images/
18. No, do not use ChatGPT to write code for you. It has no clue about what the code actually does or what you want it to do. https://www.kdnuggets.com/5-coding-tasks-chatgpt-cant-do 
19. A strategic look at how to integrate AI chatbots into your website: https://www.datasciencecentral.com/unleashing-innovation-how-ai-chatbots-transform-your-website-strategy/
20. A detailed overview of the copyright lawsuits around AI. But, it's not AI infringing on human rights, but the people using them. Otherwise its like saying that hammers have it in for my left thumb, rather than me being clumsy: https://dataconomy.com/2024/01/08/ai-copyright-lawsuits/