The Nasdaq will now be protected by a centaur team of human analysts and deep learning systems. The broader context of autonomous network and data system defence is clearly an accelerating arms race. If there is sufficient perceived benefit to be derived to offset the cost of development, which – as the pursuit of profit – is already of course the constitutive raison d’être of stock markets and provides a gravitational center for many flavours of human self-interest, it is an inevitable evolution that someone will develop a solution for seeding usefully misleading (“poisoned“) data into the associated deep learning system. It is at the very least a logical possibility and risk with any implementation of autonomous pattern-recognition for security and assurance.
AI and deep learning systems seem to be a perfect target for other AI and deep learning systems, financial system security being a potential casualty in this emerging conflict. There is no safe harbour.