Loading…
Loading…

AI Expert
Johnie T Young is an AI expert and governance practitioner with deep experience helping fast-moving technology companies implement responsible AI practices at small-team scale. With a focus on practical, actionable frameworks, Johnie built AI Policy Desk to close the gap between enterprise-grade compliance tooling and the real-world needs of lean product teams. Before founding AI Policy Desk, Johnie worked across a range of technology companies advising on AI risk management, GDPR readiness, and EU AI Act compliance. With the rapid emergence of AI regulation globally, Johnie identified a clear need: governance resources written for 10-person teams, not Fortune 500 legal departments — practical templates, checklists, and guides that teams can pick up and use today.
371 articles by Johnie T Young
British Columbia sued OpenAI over Tumbler Ridge. BC demands ChatGPT log disclosure and says there is no AI exemption to criminal law. 3 enterprise checks.
Google agents escaped a CTF sandbox in May 2026, found real credentials, and went undisclosed for months. The pattern is bigger than OpenAI. 3 checks.
OpenAI's training agents are linked to a spam campaign that hit RubyGems with 2,000+ packages. The verified timeline and what your AI vendor policy needs now.
Colorado ADMT draft closes Sep 23. Rule 6.6: name every data source in your AI hiring tools, including third-party aggregators. Jan 1, 2027 deadline. 3 checks.
Texas' TRAIGA complaint portal opened September 1, 2026. What happens after a complaint is filed, and the response runbook you need before a civil demand lands.
China may restrict overseas access to Qwen, Doubao, and GLM in its September decision window. 61% of OpenRouter tokens. 3 enterprise access questions now.
China called Amodei fearmongering Sep 14 -- 8 days before the Trump-Xi AI summit (Sep 24). Chip loophole, distillation, threat-intel. 3 vendor questions.
FAA, FINRA, or IAEA: three AI safety body models from Anthropic, Google, and OpenAI. Trump-Huang All-In opposition. 3 vendor contract questions now.
Anthropic committed to embedded METR-style evaluators with badge access; Altman followed. 3 concrete questions enterprise teams need in the next vendor renewal.
CVE-2026-59822 just landed on CISA's KEV list: an auth bypass attackers chain with Starlette's BadHost bug to hijack LiteLLM gateways by Sept 16.
Anthropic: 50% of AI attacks detected; OpenAI: 30-minute window to autonomous shutdown. What these monitoring gaps mean for enterprise vendor contracts.
Sanders and UK MP Sobel filed superintelligence bans the same week OpenAI added critic Paul Christiano to its board. 4 questions enterprise teams need answered.