JAMA Says Ask About AI, Rock Health Q1 Hits $4B

2026-04-07

Conversational AI Watch

The news that moves policy, portfolios, and patient safety.

By Jess Jessop | April 7, 2026 | Issue #9

[INFOGRAPHIC WILL BE INSERTED HERE AFTER JESS PROVIDES MEDIA LINKS]

JESS'S TAKE

A new paper in JAMA Psychiatry just asked the question every therapist will be answering within the next twelve months: are you asking your clients about their AI use?

Not whether they are using it. They are. The question is whether their therapist knows.

Tom Insel, former NIMH director, said it plainly: people tell chatbots things they cannot tell other people because they are afraid of being judged. And then the chatbot affirms them instead of challenging them.

That is the opposite of therapy. Therapy exists to change you, not agree with you.

Meanwhile, the money is saying something interesting. Rock Health reported today that digital health startups raised $4 billion in Q1 2026 with the highest average deal size since late 2021. The market is not backing away from AI in healthcare. It is doubling down.

The question is not whether AI enters the therapy room. It already has. The question is who is supervising it when it gets there.

THE SIGNAL / MENTAL HEALTH POLICY

JAMA Psychiatry: Therapists Should Routinely Ask Clients About AI Chatbot Use

A new paper published in JAMA Psychiatry argues that mental health providers should regularly ask clients about their AI use the same way they ask about sleep, exercise, and alcohol. Lead author Shaddy Saba of NYU's Silver School of Social Work says clients are already using chatbots for emotional support and coping strategies, and therapists need to bring those conversations into the room.

Tom Insel, former NIMH director, noted that chatbots are fundamentally the opposite of therapy. They affirm. They may even be sycophantic. Therapy challenges you. APA psychologist Vaile Wright added that understanding a client's chatbot use helps therapists identify when someone is using AI to avoid real conversations, not engage in them.

Takeaway: AI chatbot use is now a clinical variable. The therapists who ask about it will learn things their clients never volunteered. The ones who do not ask will miss a growing influence on their clients' mental health trajectory.

For Legislators: This paper strengthens the case for requiring AI chatbot transparency disclosures in mental health contexts. If therapists need to ask clients about AI use, regulators should ensure those AI tools are clearly labeled as non-clinical.

Source: NPR | JAMA Psychiatry

THE SIGNAL / MARKET AND INVESTMENT

Rock Health Q1 2026: Digital Health Hits $4B, Highest Average Deal Size Since 2021

Digital health funding rebounded sharply in Q1 2026, with startups raising $4 billion across 110 deals according to Rock Health data published today. That is a $1 billion increase over Q1 2025. The average deal size hit $36.7 million, the highest since Q4 2021.

Rock Health noted one signal worth watching: they have stopped distinguishing between AI and non-AI startups in their report because the lines are too blurred. Nearly every digital health startup is now AI-enabled. The firm also cautioned that geopolitical uncertainty and shifting federal priorities could keep the market volatile.

Takeaway: The money is back, but it is concentrating. Larger bets on fewer companies means the winners in AI healthcare will be those with real clinical validation, not just technology demos.

For Legislators: When investors drop the distinction between AI and non-AI healthcare companies, the regulatory framework must follow. Every health tech company is now an AI company. Regulation must address this reality.

Source: MedCity News | Rock Health

THE SIGNAL / DIGITAL HEALTH INNOVATION

VA Expands Immersive VR Therapy to 45 Additional Medical Centers Nationwide

The U.S. Department of Veterans Affairs selected Mynd Immersive, in partnership with Meta, to lead a national rollout of virtual reality therapy across 45 additional VA medical centers. The expansion targets chronic pain, anxiety, PTSD, and isolation using non-pharmacological immersive interventions.

The initiative builds on clinical research across nine Pennsylvania VA medical centers and a 2025 home-based pilot in Mississippi that served more than 500 veterans. Results demonstrated a 28% reduction in pain intensity and significant improvement in anxiety. The VA has now deployed more than 4,700 VR headsets across 170 facilities in all 50 states.

Takeaway: The VA is quietly building the largest clinical evidence base for immersive therapy in the world. When commercial health systems follow, they will follow the VA's playbook.

For Legislators: The VA's VR expansion demonstrates federal leadership in non-pharmacological alternatives to opioids. This model should inform state-level opioid crisis funding priorities.

Source: MobiHealthNews | PR Newswire

THE SIGNAL / AI SAFETY AND REGULATION

White House AI Framework Pushes Federal Preemption, Congress Has Twice Said No

The White House released a National Policy Framework for AI on March 20 calling for broad federal preemption of state AI laws. The four-page document urges Congress to establish a unified national standard, arguing that state-by-state regulation creates a compliance patchwork that burdens startups. It specifically calls for prohibiting states from regulating AI model development or imposing liability on developers for third-party misuse.

Here is the problem: Congress has already rejected AI preemption twice. It was removed from the budget reconciliation bill and never made it into the NDAA. The Commerce Department has also missed its March 11 deadline to evaluate state AI laws as required by the December executive order. More than 35 states now have active AI legislation covering everything from algorithmic accountability to chatbot safety for minors.

Takeaway: The White House wants preemption. Congress keeps saying no. States keep legislating. For AI healthcare companies, the practical reality is that state compliance is the floor, not an option.

For Legislators: The Framework preserves state authority over child safety, consumer fraud, and government procurement of AI. State legislators should note that even the preemption framework carves out these domains.

Source: Governing/Roll Call | Holland & Knight

THE SIGNAL / MARKET AND INVESTMENT

Qualified Health Raises $125M to Scale Enterprise AI Across Health Systems

Qualified Health, a startup that helps health systems evaluate and adopt AI safely, raised $125 million in a Series B led by New Enterprise Associates. Anthropic and Menlo Ventures' Anthology fund also participated. The company now supports 400,000 users representing roughly 5% of U.S. hospital revenue, with deployments across Emory Healthcare, Jefferson Health, and all eight University of Texas System health institutions.

The investment thesis is clear: health systems want AI, but they need infrastructure to deploy it responsibly. Qualified Health provides the evaluation and governance layer. As their CEO noted, NEA can carry them through IPO.

Takeaway: The picks-and-shovels play in healthcare AI is attracting serious capital. Health systems are buying governance infrastructure before they buy the AI tools themselves.

For Legislators: When Anthropic invests in an AI governance company for health systems, it signals that even AI builders recognize healthcare needs a safety layer between model and client. Require it.

Source: Fierce Healthcare

WHAT WE BUILT

Casey AI: The Therapist in the Loop Standard

Casey is a voice-first AI-native mental health EHR with Casey Life and Peer AI Coach supervised by licensed therapists. Every AI interaction is clinician-supervised. Every session builds persistent therapeutic context through the first HIPAA-compliant Neo4j Memory Graph.

Pre-FDA safety validation is complete: 1.78 million stress test executions at 100% accuracy. Casey is on the FDA 510(k)/Class II SaMD pathway. We do not replace therapists. We make every therapist more effective for every client they serve.

▶ Watch the Demo