Publication date:
Publication from:
From July 20 to 22, CONTIO Tech Service Delivery Director Andrew Kovalyov attended TrustCon 2026 in San Francisco, an annual international conference that brings together experts from the technology industry, the public sector, academia, and civil society organizations. TrustCon is a leading event in Trust & Safety, covering the protection of online platforms and communities, privacy, content moderation, fraud prevention, cybersecurity, and efforts to combat abusive behavior and disinformation. Throughout three days of sessions and conversations, one question kept coming up: how can we protect users when rules offer no clear answer and new risks emerge faster than guidelines can be updated?
Trust & Safety has long extended beyond social media. A system may fail to detect a dangerous scenario, a fraudster may impersonate someone else, and a legitimate post may appear to violate the rules when viewed without context. At TrustCon, Andrii Kovalyov took part in discussions that examined these risks from an operational perspective. How can rules and technology be turned into safeguards that work every day?
AI systems must be tested for potential abuse before users gain access to them. A discussion with Microsoft experts focused on red teaming – identifying model vulnerabilities before release. At the Human-in-the-Loop roundtable, the focus shifted to the role of human judgment: teams must be able to stop or correct an automated decision, not merely observe it.
Once a service goes live, it is not enough to identify a risk – it must also be handled correctly. Conversations with representatives from Helix and BATLR.ai covered voice-based identity verification and AI tools for detecting fraud across marketplaces and online communities. Technology only flags a potential threat. The team must then review the signal, assess the risk, and decide how to respond in a way that protects users.
Often, a single rule is not enough to understand the context. That is where some of the most important and interesting discussions begin. For example, conversations with representatives from the Center for Democracy & Technology and Twitch explored the complex roles of political content creators. The same person may be a journalist, activist, advertiser, or representative of a political campaign. This is why Trust & Safety is about far more than policies. Above all, it requires a well-prepared team and a clear course of action for ambiguous situations.
Trust & Safety teams can rarely wait for complete certainty. That is why businesses need to define in advance how high-risk signals will be reviewed, when decisions will be escalated, and when human intervention is required. This preparedness is what turns individual technologies into a reliable operational system,
said Andrew Kovalyov, Service Delivery Director at CONTIO Tech.
TrustCon may be over, but the most practical work begins after returning home – when insights are put into practice by teams. Continuing this conversation within the IT Ukraine Association is essential, as marketplaces, fintech, game development, service providers, and AI companies all share the same challenge: identifying a problem before it becomes an incident.
