<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/9385770ff2e7466cb457e0d5f1cd49a3&quot; frameborder=&quot;0&quot; width=&quot;2560&quot; height=&quot;1920&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1920</height><width>2560</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1920</thumbnail_height><thumbnail_width>2560</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/9385770ff2e7466cb457e0d5f1cd49a3-5853276f6cab4628.gif</thumbnail_url><duration>70.583333</duration><title>Vlad and Sonya on building a culture of security in workflows</title><description>In this video, I discuss the importance of shifting AI security practices to the left, meaning we should teach secure usage of AI as soon as it is integrated into workflows, rather than waiting until the end of production. By identifying vulnerabilities early through threat modeling and user cases, we can avoid costly fixes later in the process. Additionally, I emphasize the need to foster a security culture within our organization by having security champions in various business units. It&apos;s crucial that we view security as an enabler rather than a blocker, as we are all working towards the same goals. I encourage everyone to actively participate in promoting this security mindset.</description></oembed>