This current discussion if AI can end humanity worries me. AI operators like OpenAI need to stop describing their models like gifted children with their own minds. AI models don't have a conscience, they follow paths learned from human behavior, both good and bad. Just faster....

I have written about what AI detection and watermarking are doing to the world of writing. It includes a behavioral analysis of the currently best detector, Pangram and boils down to recognizing that all detection ultimately measures the wrong thing. Link here: www.9senses.ai/the-broken-gate

AI detection: The broken gate - 9senses.ai

AI detectors and the new Claude watermark measure contact with AI, and are easy to evade. Honest AI users get marked; those who really cheat walk easily. We discuss why detection is the wrong approach, and what could actually work.

9senses.ai

How does it sound when AI is singing a ballad about AI? Listen and be spooked....

A friend's daughter spent three weeks on a high school assignment. No AI, solid research, a great presentation she could defend. She received 14 out of 15 points. A result to be proud of. Some peers using AI got 15 points. Our education needs an overhaul: www.9senses.ai/the-lost-generation

The Lost Generation: How AI Disrupts Learning | 9senses

AI now does the work that turns beginners into experts. What that means for students, graduates, universities and companies, and what to do about it.

9senses.ai

Why does an AI skeptic set up an AI transformation consultancy? Many people have asked me this. Simple answer: Because AI is here if we want it or not, and ignoring or simply saying "no" doesn't work for any business, because AI creeps in through third-party tools and employees who sneakily use it.

AI fuels the "rich get richer" paradigm more than anything else. If you don't interact with AI in English, you're at a disadvantage by default. And building AI solutions is more complex too. Read our thoughts on the issue: www.9senses.ai/lost-in-tran... #MultilingualAI #AIadoption

Why AI Underperforms In Non-English Languages | 9senses

Over 40% of AI training data is English. What that means for German, Hindi and every other language, with the numbers, and what good non-English AI takes.

9senses.ai

Some AI experts warn that AI could lead us into an apocalypse. I disagree wholeheartedly. Apocalypse becomes possible only if we remove ourselves from decisions that still require human judgment and accountability. www.9senses.ai/can-ai-end-h...

Can AI End Humanity? The Real Risk Is Us | 9senses

Dario Amodei puts the odds of catastrophe at 25%. Why the danger isn't machines revolting but humans outsourcing judgment, from hospitals to the battlefield.

9senses.ai

AI is already inside companies and education systems, whether we want it or not. Our role is to decide where it belongs - and what must remain human. The risks of enthusiastically embracing AI are as bad as refusing to acknowledge the above. We all have to become "AI Wise."