SEALISON FOR HEALTHCARE · PATIENT-FACING AI
In healthcare, “almost correct” is not enough.
The problem is not just what the AI system computed. It is what the patient, member, or operator actually saw — and you may need to prove that later.
REALITY CHECK
PATIENT SAYS
“Your AI told me I was eligible for this treatment.”
YOUR SYSTEM SAYS
“Our records show a different interpretation.”
This is no longer a customer support issue. It’s a medico-legal question.
WHAT TEAMS RELY ON TODAY
EHR system logs
Audit trail exists but entries remain editable by privileged users. No cryptographic guarantee.
Portal message history
Can be retroactively changed. No independent timestamp. No tamper-evidence.
Screenshots from patients
Not reliable evidence in medical litigation. Trivially modified. No cryptographic anchor.
When a patient or their lawyer asks for proof of what was shown — “we have logs” is the beginning of a problem, not an answer.
WITH SEALISON
01
Hash
The exact AI output is hashed with SHA-256 at the moment it's produced.
02
Seal
The hash is sealed in a signed, append-only chain. Ed25519 signatures. Any change is cryptographically detectable.
03
Keep
The original content stays with you. SEALISON does not store what your AI said — only the proof.
04
Verify
Later, you (or anyone) can reproduce the interaction and verify it matches. Public CLI verifier planned for stronger independent verification.
EU AI ACT · HEALTHCARE = HIGH RISK
EU AI Act obligations for healthcare AI systems:
Sealison does not determine whether an AI system is high-risk and does not provide legal compliance certification. It provides cryptographic evidence that can support auditability and compliance demonstrations.
Not “screenshots in the chart.” Cryptographic evidence.
WHERE THIS APPLIES
Eligibility explanations
AI tells patients what treatments or benefits they qualify for. Later disputed by insurers or patients.
Coverage & reimbursement
AI explains what will or won’t be paid. Medico-legal consequences when disputed.
Clinical-adjacent recommendations
AI suggests next steps, specialist referrals, lifestyle changes. Must be auditable.
Patient-facing treatment options
AI presents options, explains risks. Patients challenge what they were told.
WHAT CHANGES
BEFORE
“We believe the AI said this.”
EHR logs. Hopefully intact. Medico-legally contestable.
AFTER
“We can prove the AI said this.”
Cryptographic proof. Verifiable by any third party. Independently verifiable.
TRY IT NOW
Generate one from scratch, or verify one we already published. Takes 30 seconds.
Working with AI systems that talk to customers? We can show you how SEALISON fits in. 10 minutes. You test it yourself.
Or email us directly at [email protected]