Ep104: Partnership in Innovation - How ActiveFence and AWS are De-risking AI

Noam Schwartz introduces himself as ActiveFence CEO

Former intelligence officer specializing in open source intelligence

Mission: protect online experiences for everyone everywhere

Online platforms constantly hammered by various attacks

Attacks include cybersecurity, abuse, hate speech, spam

Companies playing endless whack-a-mole game with violations

Need scalable solution that works across languages/formats

Developed enterprise-grade technology for sophisticated companies

Amazon became customer and great partner early on

Generative AI introduction changed the game completely

LLMs non-deterministic unlike traditional programmed chatbots

Same input produces different outputs each time

AI deployed in customer support, healthcare, airlines

New risks when models speak on company's behalf

One bad output creates legal and reputational damage

Companies need to deploy public-facing AI safely

Transition affects healthcare, finance, gaming, government sectors

Building on years of user-generated content expertise

No specific ChatGPT moment triggered their AI pivot

ActiveFence was AI company since day one

Model companies like Amazon, Nvidia asked for help

Realized their expertise perfectly suited for AI safety

Staying on top of AI developments is impossible

Focus on customer adoption, not every new release

Main enterprise challenge is trusting AI technology

Unrealistic expectations for 100% accuracy from AI

Most companies will license existing models, not build

Security solutions remain independent like traditional cybersecurity