The Permission Layer: Who Told the Al It Could Do That? Richy Glassberg on Data Privacy, Consent, and Al Innovation

AI agents can retrieve, combine, analyze, infer from, and act on enormous amounts of data—often at a speed no human compliance team can match. But access to data does not automatically confer the right to use it. So who sets the rules, and who remains accountable when an AI system crosses the line?In this episode of Signal & Noise, Brett House and Rio Longacre sit down with Richy Glassberg, Co-Founder and CEO of SafeGuard Privacy, for a candid and wide-ranging conversation about privacy, consent, AI governance, and the digital advertising industry’s long history of creating problems it later asks technology to solve.Richy brings a rare perspective to the discussion. He helped build CNN.com’s commercial business, co-founded the IAB, worked across publishing, agencies, ad tech, and media, and now leads a company focused on making privacy compliance and vendor diligence standardized, operational, and auditable.The conversation begins with a provocative argument: AI may not require an entirely new category of privacy law because AI is ultimately software—and existing rules governing data use, discrimination, consent, and accountability still apply. The real challenge is enforcing those rules as AI dramatically increases the speed, scale, and complexity of data use.Richy explains why companies are now responsible for privacy compliance throughout their vendor chains, including the DSPs, publishers, data brokers, identity providers, models, APIs, and other partners involved in a transaction. When one black box passes data to another black box—and AI begins making decisions across the entire chain—policies and promises are no longer enough. Organizations need standardized diligence, enforceable controls, ongoing monitoring, and proof.The group also examines why today’s consent system is fundamentally broken. Cookie banners have created consent fatigue without giving consumers meaningful understanding or control. Privacy policies are rarely read, permissions do not travel cleanly across platforms, and people can opt out in one place only to reappear in the same identity graph somewhere else.Other topics include:• Why an AI agent should never have more authority than the person or organization it represents• The tension between giving AI more context and protecting individual privacy• Why human oversight remains essential in agentic systems• How marketers should assess and monitor every company handling their data• Why privacy diligence must become machine-readable for real-time agent decisions• The failure of one-to-one targeting and the industry’s obsession with questionable audience data• How poor frequency management is damaging the connected TV experience• Why better privacy practices could become a mark of data quality and competitive differentiation• The threat AI-generated content poses to trusted information and the open internet• Whether consumer-controlled data and permission agents could produce a healthier advertising ecosystemIt’s a funny, blunt, and occasionally uncomfortable conversation about what responsible data use should look like when machines can move faster than the institutions meant to govern them.Learn more about SafeGuard Privacy: https://safeguardprivacy.com/#ArtificialIntelligence #DataPrivacy #AIPrivacy #AIGovernance #Consent #DigitalAdvertising #AdTech #AgenticAI #PrivacyTech #MarketingTechnology #DataGovernance #ProgrammaticAdvertising #SignalAndNoisePodcast