![How to Create Your Own AI Toolkit with Taylor Radey [MAICON 2025 Speaker Series]](https://www.marketingaiinstitute.com/hubfs/MAICON-Speaker_Series-Taylor.png)
![[The AI Show Episode 154]: AI Answers: The Future of AI Agents at Work, Building an AI Roadmap, Choosing the Right Tools, & Responsible AI Use](https://www.marketingaiinstitute.com/hubfs/ep%20154%20cover.png)
Weekly Review 20 June 2025
Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):Wikipedia kills AI summaries of its pages: https://www.extremetech.com/computing/wikipedia-editors-revolt-over-ai-summaries Given how many bots scrape Wikipedia for training data, this is probably a good thingCorporate adoption of AI is levelling off: https://techcrunch.com/2025/06/09/corporate-ai-adoption-may-be-leveling-off-according-to-ramp-data/Sam Altman is very smart but he's no scientist. Current AI are not going to drive scientific discoveries anytime soon, although they will be useful to accelerate some parts of them: https://www.bigdatawire.com/2025/06/03/ai-agents-to-drive-scientific-discovery-within-a-year-altman-predicts/The thing with AI predicting your lifespan is, you'll only ever know if it under-estimated: https://www.nzherald.co.nz/world/how-many-years-do-i-have-left-an-app-gave-me-some-helpful-insights/6YFWBORXD5AI3NHGVQWQKYLGPY/Trying to automate any business operation with AI is fraught. Trying to automate a government with AI is going to be hilarious: https://www.theregister.com/2025/06/10/trump_admin_leak_government_ai_plans/There appears to be a fairly widespread use of AI to generate legal filings, and judges are getting sick of it: https://techcrunch.com/2025/06/07/lawyers-could-face-severe-penalties-for-fake-ai-generated-citations-uk-court-warns/AI in the browser seems like a good idea, but the privacy issues around it have yet to be resolved: https://www.techrepublic.com/article/news-arc-dia-the-browser-company/Retrieval-augmented Generation is helping make generative AI easier to roll out: https://www.techrepublic.com/article/news-gartner-generative-ai-apps-data-management-platforms/Generative AI suck at chess: https://www.extremetech.com/computing/chatgpt-lost-a-chess-game-to-an-atari-2600AI are the definition of garbage in, garbage out. Proper data annotation is essential: https://dataconomy.com/2025/06/12/why-data-labeling-quality-makes-or-breaks-ai-models/Now is not a good time to be demolishing guardrails on AI: https://www.theverge.com/politics/681727/ro-khanna-ai-state-law-moratorium-reconciliationAn AI that can interpret genomic data, not just scan for patterns: https://dataconomy.com/2025/06/10/this-ai-explains-your-genes-the-way-a-doctor-would/To be clear, this isn't AI restoring old artworks, it is AI guessing what it could have looked like: https://www.theguardian.com/science/2025/jun/11/researchers-create-ai-based-tool-that-restores-age-damaged-artworks-in-hoursAI have limitations in their reasoning, and these limitations appear to be fundamental: https://www.stuff.co.nz/world-news/360719320/new-research-offers-fresh-warnings-about-ability-aiConstant testing of AI has a cost, but it lowers costs in the long run: https://www.theregister.com/2025/06/08/chatterbox_labs_ai_adoption/AI is being used in recruitment, but people aren't happy about it: https://www.techrepublic.com/article/ai-and-recruitment/That's one way of stopping students using AI to cheat: https://www.theverge.com/news/682737/china-shuts-down-ai-chatbots-exam-seasonAI might currently be used as a screening tool in recruitment, but that doesn't mean that any occult biases in the AI won't inappropriately screen people out: https://www.computerworld.com/article/4005351/surprise-employers-are-using-ai-to-interview-you.htmlThe connection between AI and galactic colonisation seems a bit tenuous to me: https://futurism.com/google-deepmind-hassabisSome policies to mitigate the risk of shadow AI in the workplace: https://devclass.com/2025/06/10/shadow-ai-in-the-enterprise-managing-risk-without-slowing-progress/How AI is contributing to the fight against cancer: https://dataconomy.com/2025/06/10/how-ai-is-fueling-our-hope-in-the-fight-against-cancer/Some strategies to catch the errors made by generative AI: https://www.informationweek.com/machine-learning-ai/how-to-successfully-catch-generative-ai-errorsWhile it's not the only application of robust quantum computers, AI will certainly benefit from this technology: https://www.techrepublic.com/article/news-ibm-quantum-computer-error-correction-2029/How DevOps can accelerate rollout of AI: https://www.datasciencecentral.com/unlocking-the-value-of-ai-with-devops-accelerated-mlops/
Some interesting links that I Tweeted about in the last week (I also post these on Mastodon, Threads, Newsmast, and Bluesky):
- Wikipedia kills AI summaries of its pages: https://www.extremetech.com/computing/wikipedia-editors-revolt-over-ai-summaries Given how many bots scrape Wikipedia for training data, this is probably a good thing
- Corporate adoption of AI is levelling off: https://techcrunch.com/2025/06/09/corporate-ai-adoption-may-be-leveling-off-according-to-ramp-data/
- Sam Altman is very smart but he's no scientist. Current AI are not going to drive scientific discoveries anytime soon, although they will be useful to accelerate some parts of them: https://www.bigdatawire.com/2025/06/03/ai-agents-to-drive-scientific-discovery-within-a-year-altman-predicts/
- The thing with AI predicting your lifespan is, you'll only ever know if it under-estimated: https://www.nzherald.co.nz/world/how-many-years-do-i-have-left-an-app-gave-me-some-helpful-insights/6YFWBORXD5AI3NHGVQWQKYLGPY/
- Trying to automate any business operation with AI is fraught. Trying to automate a government with AI is going to be hilarious: https://www.theregister.com/2025/06/10/trump_admin_leak_government_ai_plans/
- There appears to be a fairly widespread use of AI to generate legal filings, and judges are getting sick of it: https://techcrunch.com/2025/06/07/lawyers-could-face-severe-penalties-for-fake-ai-generated-citations-uk-court-warns/
- AI in the browser seems like a good idea, but the privacy issues around it have yet to be resolved: https://www.techrepublic.com/article/news-arc-dia-the-browser-company/
- Retrieval-augmented Generation is helping make generative AI easier to roll out: https://www.techrepublic.com/article/news-gartner-generative-ai-apps-data-management-platforms/
- Generative AI suck at chess: https://www.extremetech.com/computing/chatgpt-lost-a-chess-game-to-an-atari-2600
- AI are the definition of garbage in, garbage out. Proper data annotation is essential: https://dataconomy.com/2025/06/12/why-data-labeling-quality-makes-or-breaks-ai-models/
- Now is not a good time to be demolishing guardrails on AI: https://www.theverge.com/politics/681727/ro-khanna-ai-state-law-moratorium-reconciliation
- An AI that can interpret genomic data, not just scan for patterns: https://dataconomy.com/2025/06/10/this-ai-explains-your-genes-the-way-a-doctor-would/
- To be clear, this isn't AI restoring old artworks, it is AI guessing what it could have looked like: https://www.theguardian.com/science/2025/jun/11/researchers-create-ai-based-tool-that-restores-age-damaged-artworks-in-hours
- AI have limitations in their reasoning, and these limitations appear to be fundamental: https://www.stuff.co.nz/world-news/360719320/new-research-offers-fresh-warnings-about-ability-ai
- Constant testing of AI has a cost, but it lowers costs in the long run: https://www.theregister.com/2025/06/08/chatterbox_labs_ai_adoption/
- AI is being used in recruitment, but people aren't happy about it: https://www.techrepublic.com/article/ai-and-recruitment/
- That's one way of stopping students using AI to cheat: https://www.theverge.com/news/682737/china-shuts-down-ai-chatbots-exam-season
- AI might currently be used as a screening tool in recruitment, but that doesn't mean that any occult biases in the AI won't inappropriately screen people out: https://www.computerworld.com/article/4005351/surprise-employers-are-using-ai-to-interview-you.html
- The connection between AI and galactic colonisation seems a bit tenuous to me: https://futurism.com/google-deepmind-hassabis
- Some policies to mitigate the risk of shadow AI in the workplace: https://devclass.com/2025/06/10/shadow-ai-in-the-enterprise-managing-risk-without-slowing-progress/
- How AI is contributing to the fight against cancer: https://dataconomy.com/2025/06/10/how-ai-is-fueling-our-hope-in-the-fight-against-cancer/
- Some strategies to catch the errors made by generative AI: https://www.informationweek.com/machine-learning-ai/how-to-successfully-catch-generative-ai-errors
- While it's not the only application of robust quantum computers, AI will certainly benefit from this technology: https://www.techrepublic.com/article/news-ibm-quantum-computer-error-correction-2029/
- How DevOps can accelerate rollout of AI: https://www.datasciencecentral.com/unlocking-the-value-of-ai-with-devops-accelerated-mlops/
