📍 Continue the ViVE 2026 conversation with GLACIS →
Healthcare AI

ViVE 2026: HealthCare AI gets asked for its receipts

At ViVE 2026, healthcare AI buyers pressed the accountability question: beyond what a model can do, what can a vendor show about how it behaved?

3 min read
Joe Braidwood
Joe Braidwood
Co-founder & CEO
3 min read

Ahead of ViVE 2026, the official conference site described the event as bringing together more than 9,000 digital health leaders in Los Angeles. Our focus going in was narrower: healthcare AI buyers were starting to ask what a vendor could show about how a system behaved.

The ViVE moment

ViVE 2026 put AI squarely into the program through its AI Zone and related agenda programming. AI was no longer a side conversation in digital health.

The sessions are polished and the demos impressive. But there’s a gap between “here’s what our AI does” and “here’s a record of what it did.”

The question nobody’s fully answering

Walk the exhibit hall and you’ll see AI promises everywhere. Ambient scribes that never hallucinate. Predictive models that flag deteriorating patients. Billing automation that catches every code.

What you won’t see: the operational record. A signed record can preserve a covered claim about which model produced an output for an encounter and what configured safety monitoring reported. Health system governance committees increasingly ask for scoped evidence alongside SOC 2 reports and guardrail descriptions; signatures support integrity and attribution, not the truth or effectiveness of every underlying claim.

This is the gap GLACIS was built to close.

“We have guardrails” is not the same as “here is a scoped record of what the configured guardrail reported for this covered event.”

What we’re bringing to the conversation

GLACIS connects configured runtime controls to signed operational records. For in-scope actions, protected payloads can stay inside the operator’s environment while bounded evidence — such as hashes, outcomes, signatures, and verification metadata — can be checked by an authorized reviewer. The record supports integrity checks for what it covers; it is not a safety certificate or a claim that every action was captured.

At ViVE, we focused on health systems navigating vendor AI procurement, digital health companies tracking state AI rules and EU AI Act obligations, and compliance teams who needed more than vendor self-attestation. Editor’s note (June 2026): Colorado law has since changed. The original Colorado AI Act (SB 24-205) was repealed and replaced by SB 26-189 (“Automated Decision-Making Technology,” signed May 14, 2026), with substantive compliance now commencing January 1, 2027.

For organizations already in buyer or assurance conversations, the consequential AI workflow assessment starts where AI already acts: what is supposed to happen, where controls block, escalate, or allow an action, and what evidence remains. It is a readiness aid, not a compliance score or framework certification.

9,000+
digital health leaders
Feb 22-25
ViVE 2026 conference dates
Jan 1, 2027
Colorado SB 26-189 (ADMT) compliance date

After the show

Joe and Jennifer attended ViVE 2026 from February 22 through 25 in Los Angeles. If you’re working through AI procurement security reviews, preparing for regulatory scrutiny, or building AI products that need to pass enterprise compliance checks, the same questions still apply after the show.

Healthcare AI buyers are moving from “trust us” to “prove it.” We’d love to discuss what that means for your organization.

Let’s continue the ViVE conversation

If you want the post-ViVE follow-up conversation, we can start with organization-wide monitoring or with one high-risk action. Either way, we map the intended rule, the control that should execute, and the evidence a reviewer would need afterward.

Talk to us

Primary sources