How will AI and cyber security interact in the future? How to make that interesting in a conference room filled with a thousand people?
by AdrianHow will AI and cyber security interact in the future? How to make that interesting in a conference room filled with a thousand people?
AI can bring many benefits. But it also increases our exposure to cyber security threats. Thinking beyond the immediate, it was a great pleasure to work with @Cécile to develop futures scenarios to push beyond our immediate concerns and think further, and then to present a shortened version of our work at the @InCyber Forum today.
In the end we focused on four areas:
In finance, it is not just about mass phishing, but about trading systems that are run by AI in the hope of gaining that extra millisecond of edge over other traders.
In democracy, it is not just about deepfake videos of politicians, but of AI knowing each voter’s profile so well that it can be used for mass astroturfing.
In war, it is not just about improving drone accuracy, but about the risk that an enemy takes humans out of the loop, so by the time our humans decide to react, our drones are already destroyed.
In health, it is not just about mass theft of personal data, but about our trusted chatbots telling us to take the wrong medication, or even deactivating pacemakers.
Regarding the room, even if by no means all people used the mentimeter, it was great to have the audience feedback on how they viewed each scenario (finance and war being seen more as elephants in the room – very plausible with big impacts – whilst democracy and health were more black swans – less likely but with big impacts), and how they thought each scenario would end (most thought the finance and health stories would end positively, but the democracy and war would end negatively…)
Taken together, this actually means the war scenario is the biggest threat – perceived as an elephant in the room, that will end badly. And no surprise, if we end up with killing machines that have no human in the loop…
Thanks to Cécile Wendling, Guillaume Tissier und Charles Citroën for their support.
If you want to see our session, use this you tube link, and take it from 1:58:00 to 2:25:00.