Darwinium的动态

“What are you going to do when your most valuable customers and your biggest threats both use the same AI bots?” Was the question I posed in my session at the MRC | Merchant Risk Council Vegas back in 2022. That time is now. In an age when it is digitally impossible to differentiate between a valid interaction and a synthetic twin the game moves from verifying identities to detecting intent. Intent detection is the common thread that links both Adversarial AI and Social Engineering attacks e.g “Are you a trusted AI bot used for evil?” or “are you a trusted user being manipulated?”. To beat AI you need better AI based on actionable intent data and requires - Breadth of data: Full view of a user journey, covering the full ‘customer perimeter’ (CDN/APIs, web and mobile) to detect hidden threats that bypass siloed solutions. - Depth of Data: Comprehensive insight and risk signaling based on the combination of device, identity, behavior, transactions and 3rd party signals. A system of record uniting security, fraud and identity. - Ability to Adapt faster than Humans: Protective measures that can hold the customers hand through any necessary friction while being able deploy new detections and models based on configuration instead of relying on engineering release cycles. Ultimately Defensive AI will need ‘human in the approval loop’ but not in the execution. I describe Darwinium as the company I have been building towards my whole life to prepare for the time when ‘machines fighting machines’ was a reality. We started building our Cyberfraud Prevention Platform over 3 years ago to address this existential threat of Adversarial AI. At that time AI was considered to be ‘fairy dust’ by fraud and security professionals. Now it’s real. #adversarialAI #journeyprotection #cyberfraud