A Seattle-based company is viewing recent developments in federal AI policy as a significant business opportunity. Reports surfaced on Wednesday indicating that the White House is developing an executive order that would instruct federal agencies to oppose or prevent AI regulations at the state level.
Joe Braidwood, an entrepreneur from Seattle, considers this news particularly advantageous for his venture. As CEO and co-founder of Glacis, a recently launched company supported by the AI2 Incubator, Braidwood is developing technology that enables businesses to demonstrate their AI safety protocols are functioning as designed. The company’s platform generates tamper-resistant documentation for each AI-powered decision, providing verification that safety mechanisms were properly executed.
According to Braidwood, the technology functions as a “flight recorder for enterprise AI.” He suggests that if the White House proceeds with directives to challenge state-level AI regulations, Glacis could evolve from an emerging startup into essential infrastructure. In a scenario where the Justice Department takes legal action against states implementing AI rules, having a neutral, cross-platform verification system could prove increasingly valuable.
The concept behind Glacis emerged from challenges related to regulatory frameworks. Braidwood, who has extensive experience in technology marketing leadership, recently closed Yara, a one-year-old venture that employed AI for mental wellness improvement. He attributed the closure to Illinois regulations that rendered AI-based therapy “effectively uninsurable.”
In a widely circulated LinkedIn post, Braidwood detailed his decision to shut down Yara while open-sourcing safety prompts he had created. He explained that the company closed after recognizing that AI became “dangerous” rather than simply inadequate when dealing with
individuals experiencing severe trauma or suicidal thoughts. This experience revealed where limitations must exist and illustrated how startups operating in high-risk AI sectors encounter insurmountable liability and regulatory challenges.
Following the post’s publication, Braidwood received responses from various professionals including regulators, medical practitioners, engineers, business founders, and insurance industry leaders. Many identified a common issue: the inability to independently confirm whether safety policies activated when AI systems make decisions.
This realization formed the foundation for Glacis. The platform creates a cryptographically signed documentation each time an AI model processes a query or executes an action, capturing the input, active safety verifications, and resulting decision. These records are immutable and generate in under 50 milliseconds. Regulators and insurance providers can authenticate these records without accessing private information. Braidwood indicates that insurers believe this capability could finally enable coverage for AI systems capable of proving rule compliance.
Dr. Jennifer Shannon, a psychiatrist and University of Washington adjunct professor, joined Braidwood as co-founder. The company is presently conducting private beta testing with digital health clients, including nVoq, while focusing on healthcare, financial technology, and insurance industries. Glacis has also joined Cloudflare’s Launchpad program.
Braidwood’s professional background includes serving as chief strategy officer at Vektor Medical, co-founding social TV platform Scener, and working as chief marketing officer at SwiftKey. Shannon brings nearly twenty years of psychiatric practice to the venture, having previously served as medical director at Cognoa and currently participating on the AI Resource Committee for the American Academy of Child and Adolescent Psychiatry.