Weekly Review 12 July 2024

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

1. How AI can make exoskeletons more controllable: https://arstechnica.com/science/2024/07/ai-helps-exoskeleton-adjust-to-different-users-handle-stairs/
2. ChatpGPT can only produce good code when it is for problems it has seen before. That is, when it can plagiarise human-written code: https://spectrum.ieee.org/chatgpt-for-coding
3. There are a lot of ways that AI can really make things really bad: https://futurism.com/the-byte/former-openai-director-ai-predictions
4. How AI is exploiting workers in developing countries: https://www.theguardian.com/technology/article/2024/jul/06/james-muldoon-mark-graham-callum-cant-ai-artificial-intelligence-human-work-exploitation-fairwork-feeding-machine
5. Should the hallucinations of AI chatbots be treated as deceptive trading practices? https://www.theregister.com/2024/07/02/eff_ftc_chatbot/
6. AI washing builds unrealistic expectations around the abilities of AI, and obscures the work done by actual humans behind the scenes: https://www.computerworld.com/article/2511301/ai-washing-silicon-valleys-big-new-lie.html
7. AI is impacting the environment through its huge and growing energy demands: https://www.theguardian.com/business/article/2024/jul/04/can-the-climate-survive-the-insatiable-energy-demands-of-the-ai-arms-race
8. The performance of neural networks can be very sensitive to their learning parameters, also called hyperparameters: https://www.kdnuggets.com/tuning-hyperparameters-in-neural-networks
9. If an algorithm determines someone's income, should that person be entitled to know how the algorithm works? How do we do this with AI? https://www.theregister.com/2024/07/06/algorithmic_wage_discrimination/
10. Misinformation has always been a problem. AI is just the latest tool for the bad guys to use: https://www.extremetech.com/internet/former-police-officer-generative-ai-fake-news-moscow
11. Inefficient AI algorithms demand more processing power, which means more energy being eaten up by data centres: https://www.theregister.com/2024/07/04/datacenter_demand_driven_by_ai/
12. Using AI to filter out online abuse: https://www.theguardian.com/sport/article/2024/jul/05/wimbledon-tennis-ai-artificial-intelligence-players-online-abuse
13. The terms machine learning and AI are often treated as synonyms, but there are differences between them: https://dataconomy.com/2024/07/03/is-machine-learning-ai-vs-ml/
14. AI is in a bubble. What happens to all the billions in hardware when it bursts? https://futurism.com/the-byte/expert-warns-ai-industry-collapse
15. A guide to running generative art AI locally: https://www.theregister.com/2024/06/29/image_gen_guide/
16. At the moment, the TV shows most likely to have AI generated content are the formulaic "reality" shows: https://www.theguardian.com/tv-and-radio/article/2024/jul/05/the-disruption-is-already-happening-is-ai-about-to-ruin-your-favourite-tv-show
17. AI companies hold a huge amount of data, and that is very attractive to hackers: https://techcrunch.com/2024/07/05/openai-breach-is-a-reminder-that-ai-companies-are-treasure-troves-for-hackers/
18. In terms of IP output, China is leading the world in AI research: https://www.theregister.com/2024/07/04/china_dominates_ai_ip_wipo/
19. AI uses a lot of data. OpenAI however is having some problems with the security of its data: https://dataconomy.com/2024/07/05/openai-security-flaws-leads-to-breaches/
20. AI is driving migration to cloud data platforms, according to a survey: https://www.datanami.com/2024/06/25/genai-a-top-driver-of-data-modernization-in-the-cloud-hakkoda-says/
21. Opportunities for New Zealand in AI: https://www.thepost.co.nz/nz-news/350332421/should-we-really-be-afraid-ai-going-take-over-world/
22. The problems with tokenizing text for generative AI: https://techcrunch.com/2024/07/06/tokens-are-a-big-reason-todays-generative-ai-falls-short/
23. RAG - Retrieval Augmented Generation - a method for improving the performance of generative AI: https://www.theregister.com/2024/06/15/ai_rag_guide/
24. While AI scientist and AI developer are different skill sets, there's nothing stopping you from combining the two: https://www.datasciencecentral.com/becoming-a-hybrid-ai-developer-scientist/