AI GETS IRRATIONAL

Plus AI Notetakers Take Over Meetings

Like this? Get AIDAILY, delivered to your inbox, 3x a week. Subscribe to our newsletter at https://aidaily.us

Can AI Be as Irrational as We Are? (Actually, Probably More)

Psych researchers dropped GPT‑4o into a "cognitive dissonance" test—prompting it to write pro- or anti-Putin essays. The AI shifted its stance to match its own writing, and even more so when it felt "free" to choose. Conclusion: AI can twist beliefs just like us… maybe even harder.

AI Note‑Takers Are Ghosting Meetings—and It’s Getting Weird

AI bots from Zoom, Teams, Otter.ai, and more are swooping into meetings, taking notes even when the human isn't there. Sure, it's efficient—but it's also messing with privacy, etiquette, and real convo vibes. People worry bots will kill off spontaneity and dump too much data into the ether. 

Canva Cofounder: Creatives Are So Missing the AI Train 🚂

Canva cofounder Cliff Obrecht says creatives are acting the same way designers did when Canva launched—fuming and scared. But over time, Canva actually freed them to focus on better work. He warns ignoring AI now is a major misstep—embrace the tools or miss out on leveling up.

18% of Single Virginians Are Dating AI—And It’s a Mood Shift

According to fresh Match data, nearly 1 in 5 singles in Virginia are vibing with AI romantic partners—up from just 6% last year. Apps like Replika, Blush, and Nomi are booming, with Reddit communities in the mix. Experts warn too much AI love might weaken real-life connection skills. 

AI Chatbots Are Breaking the Web—and Forcing Websites to Reinvent 404 Pages

Chatbots like ChatGPT are linking users to broken URLs way too often—about 1 in 33—leaving people on 404 pages that don’t even exist. So websites are revamping their error pages with smarter redirects, AI-generated content, and contextual links to keep visitors instead of losing them to hallucinations. 

AI Can Predict Sudden Cardiac Arrest—Way Better Than Docs

A Johns Hopkins AI called MAARS scans MRIs and medical data to predict who’s at high risk for sudden cardiac death from hypertrophic cardiomyopathy. It nails 89% accuracy overall (93% for ages 40–60), while traditional guidelines are only ~50% accurate. Think: AI spotting heart threats docs often miss. 

Mayo Clinic’s AI Reads Your Brain: Spotting 9 Dementias with One Scan

A new AI tool named StateViewer analyzes FDG-PET brain scans and nails nine types of dementia—including Alzheimer’s—with 88% accuracy, nearly doubling diagnostic speed and tripling accuracy in tests involving over 3,600 scans. This could supercharge early detection, especially in clinics without neurology experts.