Modelsight

Three weeks until @AISecurityForum Vegas '26. Model weight security, adversarial ML, AI insider risk, all during DEF CON week. We're sponsoring, and we'll be there. Come find us: aisecurity.forum/events/vegas-a…

"Where is AI actually being used in our environment?" Most security teams can't answer that with evidence. AI usage discovery changes that. New on the blog: modelsight.ai/blog/ai-usage-…

13% of professionals say GenAI is central to their org's workflow today. 95% expect it to be within 5 years. (@thomsonreuters) The time to build your AI visibility baseline is before it becomes load-bearing, not after.

Org-level GenAI adoption: 22% Individual GenAI usage: 41% (@thomsonreuters, 2025 GenAI in Professional Services Report) The difference between those numbers is shadow AI. You can't govern what you can't see.

From @thomsonreuters' 2025 GenAI in Professional Services Report: 41% of professionals use GenAI tools at work 52% say their org has no GenAI policy 64% have had no GenAI training Usage is ahead of governance. Visibility closes the gap.

Come see us in Vegas!

We're sponsoring @AISecurityForum's Vegas event on August 6. ~200 researchers, engineers, and policymakers working on securing powerful AI systems, during DEF CON week. Applications are open: aisecurity.forum/events/vegas-a…

Shadow AI is not employees being reckless. It is the predictable result of AI being useful and security having no way to see it. What it actually means for enterprise security: modelsight.ai/what-is-shadow… #ShadowAI #AISecurity #EnterpriseAI

Sensitive data reaches AI tools constantly. Most of it happens outside of any security control. How organizations are addressing it: modelsight.ai/how-to-prevent… #DataSecurity #AISecurity #AIGovernance

Employees are using ChatGPT and GenAI tools at work. Most organizations have no idea how much — or what data is being shared. How to get visibility: modelsight.ai/how-can-organi… #AIGovernance #GenAI #SecurityLeadership

Most companies hit the same wall with AI adoption. The problem is not the technology. Full breakdown: modelsight.ai/the-ai-adoptio… #AIAdoption #EnterpriseAI

A common question from security teams: do we need a dedicated AI visibility tool, or can we extend what we already have? We laid out the honest tradeoffs. modelsight.ai/ai-visibility-… #AIGovernance #SecurityStrategy

Most security teams can describe a suspicious login. Far fewer can describe suspicious AI usage. We wrote about what it actually looks like in practice — and why it's harder to detect than people expect. modelsight.ai/what-does-susp… #ThreatDetection #AISecurity

Shadow AI gets discussed a lot. What it actually means for enterprise security teams — and what to do about it — gets discussed less. We covered both. modelsight.ai/what-is-shadow… #ShadowAI #AIGovernance