Weekly Review 18 April 2025
Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):How you can use AI to help when you are made redundant: https://www.informationweek.com/machine-learning-ai/7-ways-generative-ai-can-help-you-survive-a-layoff No mention of what to do if it is the AI that makes you redundant, though.Personalising user interfaces using AI: https://www.datasciencecentral.com/how-ai-driven-personalization-is-transforming-user-interface-design/Predicting disasters by integrating geospatial data with AI: https://dataconomy.com/2025/04/09/googles-new-ai-could-predict-disasters-before-they-hit/Is doge using AI to spy on government employees? I wouldn't be surprised: https://www.theguardian.com/us-news/ng-interactive/2025/apr/10/elon-musk-doge-spyingYes, AI can predict traffic accidents. I had a student do it for Auckland a while ago: https://www.datasciencecentral.com/can-ai-predict-traffic-accidents/AI can be told to show their reasoning, but now they are lying about it: https://arstechnica.com/ai/2025/04/researchers-concerned-to-find-ai-models-hiding-their-true-reasoning-processes/Some guidelines for how teachers can deal with students using AI inappropriately: https://www.edutopia.org/article/responding-student-ai-useRemember, if your AI system is really contractors in the Philippines, it's fraud: https://techcrunch.com/2025/04/10/fintech-founder-charged-with-fraud-after-ai-shopping-app-found-to-be-powered-by-humans-in-the-philippines/Current AI make good advisors, but humans should always be making the final decision. But people are lazy, so they probably will just do what the machine tells them: https://www.informationweek.com/machine-learning-ai/when-should-human-decision-making-overrule-ai-AI don't have value systems, they need guardrails: https://techcrunch.com/2025/04/09/mit-study-finds-that-ai-doesnt-in-fact-have-values/The security problems around AI: https://www.informationweek.com/cyber-resilience/what-are-the-biggest-blind-spots-for-cios-in-ai-security-The New Zealand military is allowed to use killer drones, but AI can't make the kill decisions: https://www.rnz.co.nz/news/national/557573/there-should-always-be-a-human-in-charge-defence-minister-on-nzdf-s-killer-dronesFixing the robots.txt protocol to keep AI scraper bots out: https://www.theregister.com/2025/04/09/ietf_ai_preferences_working_group/As more AI incorporate reasoning into their models, benchmarking them becomes more expensive: https://techcrunch.com/2025/04/10/the-rise-of-ai-reasoning-models-is-making-benchmarking-more-expensive/Automating marking of high school writing assessments using AI: https://www.rnz.co.nz/news/national/557671/artificial-intelligence-exam-marking-on-the-way-for-year-10-writing-testsAI can't replace barristers, but they can help them out: https://www.stuff.co.nz/business/360641260/ai-having-high-performing-junior-lawyer-says-leading-barristerMeta's benchmarks for its new AI are misleading. Colour me surprised: https://techcrunch.com/2025/04/06/metas-benchmarks-for-its-new-ai-models-are-a-bit-misleading/Ways in which AI is being used in financial planning: https://www.bigdatawire.com/2025/04/08/how-ai-and-ml-will-change-financial-planning/If they're going to use AI to regrant benefits, they need to thoroughly check their training data for bias first: https://www.rnz.co.nz/news/political/557600/ministry-of-social-development-to-use-basic-ai-to-check-if-jobseekers-have-met-their-obligationsDo we just have to learn to live with hallucinations from AI? https://www.bigdatawire.com/2025/04/07/can-we-learn-to-live-with-ai-hallucinations/Using AI to raise money for charity: https://techcrunch.com/2025/04/08/a-nonprofit-is-using-ai-agents-to-raise-money-for-charity/As of 2025, AI are getting smarter, more money is being spent on them, and people are worrying more about them: https://www.techrepublic.com/article/news-stanford-2025-ai-index/AI might be cool, but don't try to use it in court, judges don't like it: https://www.theverge.com/news/646372/ai-lawyer-artificial-avatar-new-york-court-case-videoAn overview of transformers, the basis of generative AI: https://www.kdnuggets.com/wtf-is-a-transformer
Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):
- How you can use AI to help when you are made redundant: https://www.informationweek.com/machine-learning-ai/7-ways-generative-ai-can-help-you-survive-a-layoff No mention of what to do if it is the AI that makes you redundant, though.
- Personalising user interfaces using AI: https://www.datasciencecentral.com/how-ai-driven-personalization-is-transforming-user-interface-design/
- Predicting disasters by integrating geospatial data with AI: https://dataconomy.com/2025/04/09/googles-new-ai-could-predict-disasters-before-they-hit/
- Is doge using AI to spy on government employees? I wouldn't be surprised: https://www.theguardian.com/us-news/ng-interactive/2025/apr/10/elon-musk-doge-spying
- Yes, AI can predict traffic accidents. I had a student do it for Auckland a while ago: https://www.datasciencecentral.com/can-ai-predict-traffic-accidents/
- AI can be told to show their reasoning, but now they are lying about it: https://arstechnica.com/ai/2025/04/researchers-concerned-to-find-ai-models-hiding-their-true-reasoning-processes/
- Some guidelines for how teachers can deal with students using AI inappropriately: https://www.edutopia.org/article/responding-student-ai-use
- Remember, if your AI system is really contractors in the Philippines, it's fraud: https://techcrunch.com/2025/04/10/fintech-founder-charged-with-fraud-after-ai-shopping-app-found-to-be-powered-by-humans-in-the-philippines/
- Current AI make good advisors, but humans should always be making the final decision. But people are lazy, so they probably will just do what the machine tells them: https://www.informationweek.com/machine-learning-ai/when-should-human-decision-making-overrule-ai-
- AI don't have value systems, they need guardrails: https://techcrunch.com/2025/04/09/mit-study-finds-that-ai-doesnt-in-fact-have-values/
- The security problems around AI: https://www.informationweek.com/cyber-resilience/what-are-the-biggest-blind-spots-for-cios-in-ai-security-
- The New Zealand military is allowed to use killer drones, but AI can't make the kill decisions: https://www.rnz.co.nz/news/national/557573/there-should-always-be-a-human-in-charge-defence-minister-on-nzdf-s-killer-drones
- Fixing the robots.txt protocol to keep AI scraper bots out: https://www.theregister.com/2025/04/09/ietf_ai_preferences_working_group/
- As more AI incorporate reasoning into their models, benchmarking them becomes more expensive: https://techcrunch.com/2025/04/10/the-rise-of-ai-reasoning-models-is-making-benchmarking-more-expensive/
- Automating marking of high school writing assessments using AI: https://www.rnz.co.nz/news/national/557671/artificial-intelligence-exam-marking-on-the-way-for-year-10-writing-tests
- AI can't replace barristers, but they can help them out: https://www.stuff.co.nz/business/360641260/ai-having-high-performing-junior-lawyer-says-leading-barrister
- Meta's benchmarks for its new AI are misleading. Colour me surprised: https://techcrunch.com/2025/04/06/metas-benchmarks-for-its-new-ai-models-are-a-bit-misleading/
- Ways in which AI is being used in financial planning: https://www.bigdatawire.com/2025/04/08/how-ai-and-ml-will-change-financial-planning/
- If they're going to use AI to regrant benefits, they need to thoroughly check their training data for bias first: https://www.rnz.co.nz/news/political/557600/ministry-of-social-development-to-use-basic-ai-to-check-if-jobseekers-have-met-their-obligations
- Do we just have to learn to live with hallucinations from AI? https://www.bigdatawire.com/2025/04/07/can-we-learn-to-live-with-ai-hallucinations/
- Using AI to raise money for charity: https://techcrunch.com/2025/04/08/a-nonprofit-is-using-ai-agents-to-raise-money-for-charity/
- As of 2025, AI are getting smarter, more money is being spent on them, and people are worrying more about them: https://www.techrepublic.com/article/news-stanford-2025-ai-index/
- AI might be cool, but don't try to use it in court, judges don't like it: https://www.theverge.com/news/646372/ai-lawyer-artificial-avatar-new-york-court-case-video
- An overview of transformers, the basis of generative AI: https://www.kdnuggets.com/wtf-is-a-transformer
