2025 is the year of widespread generative AI video. Six months after the wider launch of OpenAI’s Sora and Google’s Veo 2, and as the ability to create, edit and use generative AI video grows in realism and ease of access, what do journalists and the public need to know about these technologies and the continuing usage of other tools for video deepfakes?
Grounded in the ongoing global work of WITNESS, which includes work on direct training globally, running the Deepfakes Rapid Response Force, and technical and policy work globally, this presentation by Sam Gregory will provide a rapid-fire primer on:
> What are the key developments in generative AI video?
> What is the landscape of deceptive video usages, drawing on lessons learned from 2024, as well as the impact of the pervasiveness of AI for plausibly claiming authentic video is falsified?
> What is the status of technical solutions and guardrails from provenance to detection to platform action?
> How is regulation relevant: from the EU AI Act to the US Trump White House to global responses?