Weekly Review 29 March 2024

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

1. Ways in which AI can speed up drug clinical trials: https://www.nature.com/articles/d41586-024-00753-x
2. A general purpose, AI enabled industrial robot that can be taught by humans remotely: https://www.freethink.com/robots-ai/general-purpose-robot
3. An entirely predictable outcome of the "publish or perish" culture in academia-researchers using AI to churn out papers that go into paper mill journals: https://www.404media.co/scientific-journals-are-publishing-papers-with-ai-generated-text/
4. The use and misuse of Red Teaming when testing AI: https://spectrum.ieee.org/red-team-ai-llms
5. Despite what Microsoft says, we need regulations to prevent further harm from AI: https://www.technologyreview.com/2023/05/16/1073167/how-do-you-solve-a-problem-like-out-of-control-ai/
6. AI is not a panacea. If the roads are already at their maximum capacity, it doesn't matter how cleverly you manage the traffic, it's going to get jammed: https://spectrum.ieee.org/bengaluru-traffic
7. More doom from an AI expert: https://www.rnz.co.nz/national/programmes/sunday/audio/2018930431/eliezer-yudkowsky-the-ai-academic-warning
8. YouTube's tools for labelling AI generated material: https://blog.google/intl/en-in/products/platforms/how-were-helping-creators-disclose-altered-or-synthetic-content/
9. An agricultural robot that uses AI to monitor crop health: https://www.freethink.com/robots-ai/agricultural-robot
10. Like any other hot topic, there will be people who prey on others to make easy money. AI is no exception: https://www.technologyreview.com/2024/03/20/1089950/ai-influencer-douyin-wechat-suspension/
11. Problems arise when AI start to replace human participants in qualitative studies: https://www.rnz.co.nz/news/on-the-inside/512007/something-felt-off-how-ai-messed-with-our-human-research-and-what-we-learned
12. Cloaking phishing emails to get them past AI filters: https://www.informationweek.com/cyber-resilience/-conversation-overflow-cyberattacks-bypass-ai-security-to-target-execs
13. Mushroom identification with AI isn't a terribly hard problem, but it is still far from perfect enough to trust your life with: https://dataconomy.com/2024/03/20/dont-trust-ai-much-especially-when-it-comes-to-mushrooms/
14. AI voice cloning. I don't particularly like my voice, why would I want to change it to another language? https://www.freethink.com/robots-ai/voice-cloning
15. Generative AI will cause a lot of changes in businesses, but it's not there yet: https://www.computerworld.com/article/3714167/how-generative-ai-will-drive-a-foundational-shift-in-your-company.html
16. Basically, generative AI is not very detail oriented: https://techcrunch.com/2024/03/21/why-is-ai-so-bad-at-spelling/
17. I'll say it again (statistically) biased data results in biases models. This is something I teach my first-year, first-semester AI students, yet companies still seem to struggle with the concept: https://www.informationweek.com/data-management/what-can-a-cio-do-about-ai-bias-
18. Like every other technology, AI will destroy some jobs but create others. Now it's being forecast that AI will create half a billion new jobs in the next ten years: https://www.computerworld.com/article/3714447/generative-ai-to-create-a-half-billion-new-jobs-heres-why.html
19. Another AI generated work, with some real howlers: https://www.theregister.com/2024/03/20/amazon_selling_crockpot_recipe_books/
20. Improved breast cancer diagnosis with AI: https://dataconomy.com/2024/03/21/mia-breast-cancer-ai-diagnosis/
21. Deepfake porn continues to be an issue: https://www.theguardian.com/technology/2024/mar/21/celebrities-victims-of-deepfake-pornography
22. Universities are building their own AI chatbots: https://www.insidehighered.com/news/tech-innovation/artificial-intelligence/2024/03/21/universities-build-their-own-chatgpt-ai
23. As the size of AI models increases, so too does the power of hardware used to train them: https://www.datanami.com/2024/03/19/nvidia-introduces-new-blackwell-gpu-for-trillion-parameter-ai-models/
24. An overview of the EU's AI law: https://www.datasciencecentral.com/the-eus-ai-act-a-measured-approach-to-innovation-and-regulation/