How the Australian Federal Police Are Using AI to Protect the Public | Microsoft AI Tour 2024

From Microsoft AI Tour, 2024 in Sydney: Karissa Breen sits down with Ben Lamont, Chief Data Officer at the Australian Federal Police, and Helen Schneider, Commander of the ACE (Australian Centre to Counter Child Exploitation), to talk about how the AFP is using AI both to investigate crime and to counter the criminal use of AI itself. Ben walks through how AI is helping the AFP process a volume of data that's gone beyond human scale: translation, transcription, video review, and telephone intercepts, always with a human in the loop before any decision affects someone's liberty. He explains why the AFP avoids generative AI for high-risk predictions in favor of more traditional, auditable methods, and why he sees AI as a force multiplier the organization couldn't do without. Helen brings the conversation to the ACE's work on victim identification and the rising challenge of photorealistic AI-generated imagery, including how the unit is partnering with industry to manage scale and using prevention and education as the core defense. Content note: this part of the conversation touches on child exploitation material and online grooming. They also cover deepfakes and how the public can start to spot them, the AFP's Responsible Technology Committee, how they test AI models for bias using broader societal datasets rather than law enforcement data alone, and the layers of parliamentary and independent oversight that keep the agency accountable. 0:00 – Intro: Ben Lamont and Helen Schneider, AFP 0:22 – How the AFP is using AI to process data at scale 1:42 – Balancing AI predictions with human decision-making 2:16 – Assurance and hallucination risk in high-stakes AI use 3:26 – AI as a "force multiplier" for the AFP 4:02 – Helen Schneider: the ACE and rising case volumes 5:00 – The victim identification challenge with AI-generated imagery 6:55 – Managing deepfakes: how the public can spot them 9:22 – Prevention, education, and awareness as core strategy 9:46 – Countering criminal early adoption of AI 12:12 – Ethical considerations in high-stakes AI use 14:36 – Governance: the AFP's Responsible Technology Committee 15:33 – Managing bias in AI models 17:11 – Reducing hallucinations through broader datasets and oversight 19:09 – Oversight: parliamentary committees and accountability 19:57 – Closing thoughts from Ben and Helen #KBIMedia #MicrosoftAI #CyberSecurity #ArtificialIntelligence #AFP #LawEnforcement #ResponsibleAI #ChildSafety #DeepfakeDetection #DataGovernance #AIEthics