As usual, I fully agree with everything you say. As the European Commission seems to be saying on how to deal with APP fraud in the context of PSD2 revision: so much more is possible for banks, using machine learning/AI. A little incentive to do a better job here in the shape of a bit more liability seems to be appropriate. See their recent PSD3/PSR proposals. Of course all this is easier said than done, and time is working against the banks I believe. GenAI developments are taking place at an incredible pace, and fraudsters will rather sooner than later be successful in deep faking biometrics and thus bypassing security (in fact, they already are to a certain extent). "Yes, but is stil early days, you can demonstrate it in a controlled lab environment, and it will take time for criminals to produce at industry scale." Reminds me of the research by Ross Anderson in the first decade of this century and the bank's reflexes, see e.g. https://krebsonsecurity.com/2012/09/researchers-chip-and-pin-enables-chip-and-skim/. I'm scared s***tless if I try to image what AI in the wrong hands can do. Not only in payments, by the way. And yes, I know, it will help finding cures for cancer too. That is the terrible dilemma here: is there still time to balance the tech optimist creativity push (sort of) renegade train to the AI goldrush (what a great stunt scene in "Mission Impossible: Dead Reckoning"!!) and some thinking about how to mitigate the inevitable negative consequences? Rhetorical question I guess; hope to see the happy end in 2024 in Part II when Ethan brings The Entity under control .....
As usual, I fully agree with everything you say. As the European Commission seems to be saying on how to deal with APP fraud in the context of PSD2 revision: so much more is possible for banks, using machine learning/AI. A little incentive to do a better job here in the shape of a bit more liability seems to be appropriate. See their recent PSD3/PSR proposals. Of course all this is easier said than done, and time is working against the banks I believe. GenAI developments are taking place at an incredible pace, and fraudsters will rather sooner than later be successful in deep faking biometrics and thus bypassing security (in fact, they already are to a certain extent). "Yes, but is stil early days, you can demonstrate it in a controlled lab environment, and it will take time for criminals to produce at industry scale." Reminds me of the research by Ross Anderson in the first decade of this century and the bank's reflexes, see e.g. https://krebsonsecurity.com/2012/09/researchers-chip-and-pin-enables-chip-and-skim/. I'm scared s***tless if I try to image what AI in the wrong hands can do. Not only in payments, by the way. And yes, I know, it will help finding cures for cancer too. That is the terrible dilemma here: is there still time to balance the tech optimist creativity push (sort of) renegade train to the AI goldrush (what a great stunt scene in "Mission Impossible: Dead Reckoning"!!) and some thinking about how to mitigate the inevitable negative consequences? Rhetorical question I guess; hope to see the happy end in 2024 in Part II when Ethan brings The Entity under control .....