One thing I still don’t fully understand: since statistical and mathematical methods have long underpinned decision-making in sectors like banking (e.g. home loans) and insurance, and those same methods form the basis of modern AI, why does the EU AI Act impose stricter regulatory thresholds now? Is it simply because AI systems can perform these decisions in a more automated or autonomous way?
WELL, great questions here too! I replied to the one you had on Linkedin as well. To answer here, when it comes to "modern AI" the key things to think of are opacity, autonomy, and scale. Traditional statistical models are pretty transparent and easy to "audit" but modern AI (cannot find a better word for this now lol) has a "black box" nature, ability to learn, and when used wrong can create quite the negative impact on people and on the society, so in my practitioner opinion (not dpo/legal) the Act wants to ensure accountability and human oversight for those powerful, less predictable systems that people build every day :) and have little control over.
Also you asked on Linkedin too if this will affect the use of genAI, hell yes, and thank god. If you're using genAI for components in high risk systems, those outputs must meet the same strict compliance for safety, fairness, and explainability as anything human made. This friction and validation, ensuring responsible deployment IS VERY MUCH NEEDED. We are not even at the peak of the stuff that can be created to harm, out of greed, out of the need to become first or just simply build with AI at all costs.
I'm giving my first baby steps in the world of data science, without being an engineer or mathematician or statistician, but working (I wish I could work more and with a team) on web analytics. I'm reading "Principles of Data Science editing by PACKT"
Once again thanks, and sorry about this questions :D
Finally I have something to point to when people, after my AI Act presentation, then ask me more about High Risk requirements!
Great job making this approachable, as I know it is not an easy task.
thank you so so much for your help!
One thing I still don’t fully understand: since statistical and mathematical methods have long underpinned decision-making in sectors like banking (e.g. home loans) and insurance, and those same methods form the basis of modern AI, why does the EU AI Act impose stricter regulatory thresholds now? Is it simply because AI systems can perform these decisions in a more automated or autonomous way?
WELL, great questions here too! I replied to the one you had on Linkedin as well. To answer here, when it comes to "modern AI" the key things to think of are opacity, autonomy, and scale. Traditional statistical models are pretty transparent and easy to "audit" but modern AI (cannot find a better word for this now lol) has a "black box" nature, ability to learn, and when used wrong can create quite the negative impact on people and on the society, so in my practitioner opinion (not dpo/legal) the Act wants to ensure accountability and human oversight for those powerful, less predictable systems that people build every day :) and have little control over.
Also you asked on Linkedin too if this will affect the use of genAI, hell yes, and thank god. If you're using genAI for components in high risk systems, those outputs must meet the same strict compliance for safety, fairness, and explainability as anything human made. This friction and validation, ensuring responsible deployment IS VERY MUCH NEEDED. We are not even at the peak of the stuff that can be created to harm, out of greed, out of the need to become first or just simply build with AI at all costs.
So I guess AI will note replace us , ahahahah.
I'm giving my first baby steps in the world of data science, without being an engineer or mathematician or statistician, but working (I wish I could work more and with a team) on web analytics. I'm reading "Principles of Data Science editing by PACKT"
Once again thanks, and sorry about this questions :D
Incredible as always
thank you so much lovely, glad to see you survived reading it!