Weekly Review 24 January 2025
Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):How to switch off the AI in Windows: https://www.computerworld.com/article/3801801/avoid-ai-windows-web.htmlThere are numerous ways fraud can be committed in the supply chain. There are some ways AI can be used to combat it: https://www.datasciencecentral.com/the-role-of-ai-in-fighting-supply-chain-fraud/Using AI to create better snake antivenom: https://arstechnica.com/science/2025/01/researchers-use-ai-to-design-proteins-that-block-snake-venom-toxins/Using generative AI for fashion design: https://techcrunch.com/2025/01/13/raspberry-ai-raises-24m-from-a16z-to-accelerate-fashion-design/How do we prepare for artificial general intelligence? https://www.informationweek.com/machine-learning-ai/are-we-ready-for-artificial-general-intelligence-The weird mistakes that AI make, and how to deal with them: https://spectrum.ieee.org/ai-mistakes-schneierThe energy demands of AI are stymieing efforts to reduce greenhouse gas emissions: https://www.theregister.com/2025/01/16/ai_datacenters_putting_zero_emissions/Organisations are under-estimating the energy demands of AI: https://www.theregister.com/2025/01/15/ai_power_cooling_demands/51% of British companies plan to invest in AI rather than staff: https://www.theregister.com/2025/01/15/uk_companies_ai_report/How Kazakhstan uses AI in prisons: https://www.computerworld.com/article/3803652/kazakhstans-ai-powered-prisons-setting-a-new-standard-for-safety-and-efficiency.htmlThe security threats that come out of AI: https://www.techrepublic.com/article/5-emerging-ai-threats-australia/The security risks of cloud-based AI: https://www.informationweek.com/machine-learning-ai/addressing-the-security-risks-of-ai-in-the-cloudMeta is just one of the companies using copyrighted work to train their AI: https://techcrunch.com/2025/01/15/in-ai-copyright-case-zuckerberg-turns-to-youtube-for-his-defense/The AI behind facial recognition are powerful but easily fooled: https://www.theregister.com/2025/01/15/make_up_thwart_facial_recognition/I've never enjoyed dealing with HR, as either an applicant or manager, but I suspect I'd enjoy dealing with an AI HR agent even less: https://www.theregister.com/2025/01/15/megan_ai_recruiting_agent/An optical chip that can handle all of the operations of a deep neural network AI: https://arstechnica.com/science/2025/01/getting-an-all-optical-ai-to-handle-non-linear-math/AI needs to work for and with employees, not vice versa: https://www.informationweek.com/machine-learning-ai/ai-s-on-duty-but-i-m-the-one-staying-lateHow to handle employee's concerns about the impact of AI: https://www.techrepublic.com/article/ai-adoption-workforce-impact-strategies/SUNY requires AI and its ethical use to be added across the board: https://www.insidehighered.com/news/tech-innovation/artificial-intelligence/2025/01/16/suny-adds-ai-education-its-informationThe thing with a multi-lingual AI is that sometimes it starts reasoning in other languages. Multi-lingual humans do a similar thing: https://techcrunch.com/2025/01/14/openais-ai-reasoning-model-thinks-in-chinese-sometimes-and-no-one-really-knows-why/People seem to forget that Asimov's three laws of robotics were a literary device to drive plots. Several of his stories are how they could be worked around. Adding a 4th law won't do much: https://spectrum.ieee.org/isaac-asimov-robotics"accuracy...over 90 percent" in using AI to predict IQ. But IQ is a continuous value, how do you use percentage to assess accuracy? https://dataconomy.com/2025/01/16/cerebrum-iq-review-can-ai-predict-iq-and-cognitive-potential63558/Another CEO of an AI company who is just completely out of touch in pushing their AI-generated garbage: https://futurism.com/suno-ai-ceo-making-musicCreators continue to be opposed to the wholesale use of their work as training data for AI: https://www.theguardian.com/technology/2025/jan/14/british-novelists-criticise-government-over-ai-theft
Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):
- How to switch off the AI in Windows: https://www.computerworld.com/article/3801801/avoid-ai-windows-web.html
- There are numerous ways fraud can be committed in the supply chain. There are some ways AI can be used to combat it: https://www.datasciencecentral.com/the-role-of-ai-in-fighting-supply-chain-fraud/
- Using AI to create better snake antivenom: https://arstechnica.com/science/2025/01/researchers-use-ai-to-design-proteins-that-block-snake-venom-toxins/
- Using generative AI for fashion design: https://techcrunch.com/2025/01/13/raspberry-ai-raises-24m-from-a16z-to-accelerate-fashion-design/
- How do we prepare for artificial general intelligence? https://www.informationweek.com/machine-learning-ai/are-we-ready-for-artificial-general-intelligence-
- The weird mistakes that AI make, and how to deal with them: https://spectrum.ieee.org/ai-mistakes-schneier
- The energy demands of AI are stymieing efforts to reduce greenhouse gas emissions: https://www.theregister.com/2025/01/16/ai_datacenters_putting_zero_emissions/
- Organisations are under-estimating the energy demands of AI: https://www.theregister.com/2025/01/15/ai_power_cooling_demands/
- 51% of British companies plan to invest in AI rather than staff: https://www.theregister.com/2025/01/15/uk_companies_ai_report/
- How Kazakhstan uses AI in prisons: https://www.computerworld.com/article/3803652/kazakhstans-ai-powered-prisons-setting-a-new-standard-for-safety-and-efficiency.html
- The security threats that come out of AI: https://www.techrepublic.com/article/5-emerging-ai-threats-australia/
- The security risks of cloud-based AI: https://www.informationweek.com/machine-learning-ai/addressing-the-security-risks-of-ai-in-the-cloud
- Meta is just one of the companies using copyrighted work to train their AI: https://techcrunch.com/2025/01/15/in-ai-copyright-case-zuckerberg-turns-to-youtube-for-his-defense/
- The AI behind facial recognition are powerful but easily fooled: https://www.theregister.com/2025/01/15/make_up_thwart_facial_recognition/
- I've never enjoyed dealing with HR, as either an applicant or manager, but I suspect I'd enjoy dealing with an AI HR agent even less: https://www.theregister.com/2025/01/15/megan_ai_recruiting_agent/
- An optical chip that can handle all of the operations of a deep neural network AI: https://arstechnica.com/science/2025/01/getting-an-all-optical-ai-to-handle-non-linear-math/
- AI needs to work for and with employees, not vice versa: https://www.informationweek.com/machine-learning-ai/ai-s-on-duty-but-i-m-the-one-staying-late
- How to handle employee's concerns about the impact of AI: https://www.techrepublic.com/article/ai-adoption-workforce-impact-strategies/
- SUNY requires AI and its ethical use to be added across the board: https://www.insidehighered.com/news/tech-innovation/artificial-intelligence/2025/01/16/suny-adds-ai-education-its-information
- The thing with a multi-lingual AI is that sometimes it starts reasoning in other languages. Multi-lingual humans do a similar thing: https://techcrunch.com/2025/01/14/openais-ai-reasoning-model-thinks-in-chinese-sometimes-and-no-one-really-knows-why/
- People seem to forget that Asimov's three laws of robotics were a literary device to drive plots. Several of his stories are how they could be worked around. Adding a 4th law won't do much: https://spectrum.ieee.org/isaac-asimov-robotics
- "accuracy...over 90 percent" in using AI to predict IQ. But IQ is a continuous value, how do you use percentage to assess accuracy? https://dataconomy.com/2025/01/16/cerebrum-iq-review-can-ai-predict-iq-and-cognitive-potential63558/
- Another CEO of an AI company who is just completely out of touch in pushing their AI-generated garbage: https://futurism.com/suno-ai-ceo-making-music
- Creators continue to be opposed to the wholesale use of their work as training data for AI: https://www.theguardian.com/technology/2025/jan/14/british-novelists-criticise-government-over-ai-theft
What's Your Reaction?
