How marketplaces split the work between models and moderators to turn new risks into enforced policy.

Counterfeiters don't give up when a listing comes down—the determined ones return under a new storefront with the same stock. Across the industry, that persistence is what makes counterfeits so hard to stamp out: Takedowns treat the symptom, not the operator behind it. Anna Pogreb of G2 Risk Solutions and Josefina Arias of Amazon make the case for treating detection, enforcement, and prevention as one connected system—where every flagged listing tells you more about the seller, and every seller's history changes how you read the next listing. They'll walk through how that connection sharpens enforcement decisions, which partners make deterrence stick, and what any of us can do to raise the cost of coming back. Key takeaways - What detection can realistically see, and how much confidence you need before you act - How listing signals and seller profiles connect to reveal the operator behind the product - Where brands and law enforcement fit in, and how they turn enforcement into lasting deterrence

In the AI era, the oldest question in renting — is this person who they say they are? — has become the hardest to answer. Payslips can be generated, references fabricated, and forged documents now sail past checks that a letting agent could once catch by eye. Identity verification has become core infrastructure for the rental market, yet every check is also a moment of friction in a business built on conversion.

AI fairness is often treated as a technical problem, evaluated through model metrics and algorithmic audits. Yet bias rarely enter systems through code alone. It is introduced through human workflows, via incentives, cultural drift, uneven training, and coercive or opaque operational practices. This session argues that Trust & Safety teams now function as the last enforceable line of AI governance. While most audits focus on model outputs, they frequently ignore the global human pipelines that label data, review edge cases, and shape system behavior over time. These blind spots leave organizations exposed to hidden bias and regulatory risk. The panel explores how fairness can be operationalized inside Trust & Safety programs through workforce design, quality assurance, reviewer protection, and transparent feedback loops, shifting fairness from a claim to something that can be demonstrated and defended.

What happens when platforms stop simply presenting choices and start making or acting on them? We look at the shift from marketplace to agent from both the engineer’s and lawyer’s perspectives, and ask whether our existing ideas of platform responsibility and liability still work in the age of AI.