Immigration psychology evaluations: asylum assessments, documentation sensitivity, and vendor data custody

Immigration psychological evaluations occupy a distinct position in the landscape of mental health documentation. Unlike a therapy progress note — which records an ongoing clinical relationship — an immigration evaluation is a forensic document: a structured clinical opinion written for a legal proceeding, typically at the direction of an immigration attorney. The content of the evaluation and the clinical interview that underlies it is unusually sensitive even by mental health standards: persecution narratives identifying specific countries and persecutors, detailed trauma histories that form the evidentiary backbone of asylum claims, and expert clinical opinion on whether the client's psychological presentation is consistent with the events they described.

When a cloud AI scribe is present during the clinical interview that forms the basis of that evaluation, the vendor holds a verbatim record of all of it — captured and retained on vendor infrastructure, independently of the evaluator's own written report. This post examines what immigration psychological evaluations actually contain, what cloud AI scribes capture from these sessions, and what the data-custody structure means for one of mental health practice's most vulnerable client populations.

What immigration psychological evaluations are

Immigration psychological evaluations are forensic assessments conducted by licensed mental health professionals — typically psychologists, LCSWs, or LPCs with forensic training — at the request of an immigration attorney or, less commonly, the individual applicant directly. Unlike therapy notes, which document an ongoing therapeutic relationship, an immigration evaluation is a one-time or multi-session structured assessment intended to produce a written report submitted as evidence in an immigration proceeding.

The most common evaluation types include:

• Asylum and withholding of removal evaluations — establishing that the applicant has experienced past persecution or has a well-founded fear of future persecution on account of race, religion, nationality, membership in a particular social group, or political opinion; documenting the psychological sequelae of that persecution.
• VAWA self-petition evaluations — assessing the psychological effects of domestic violence and abuse by a US citizen or permanent resident spouse or parent, supporting Violence Against Women Act self-petitions.
• U-visa mental health evaluations — assessing substantial physical or mental abuse suffered by crime victims who cooperated with law enforcement, for applicants seeking U nonimmigrant status.
• T-visa evaluations — documenting the trauma consistent with human trafficking for individuals seeking T nonimmigrant status as trafficking victims.
• Extreme hardship evaluations — assessing the exceptional psychological impact on a qualifying US citizen or permanent resident family member if the non-citizen applicant is deported or denied admission.
• Competency evaluations in removal proceedings for respondents who may be unable to represent themselves before an immigration judge.

Each type generates a written report submitted as evidence. The clinical interview session or sessions in which the evaluator gathered information from the client are the evidentiary foundation of the expert's written opinion — and the record of what the client said, in their own words, during that interview.

What these evaluations document

The clinical interview for an immigration psychological evaluation elicits several categories of information that are unusually sensitive even in a mental health context.

The persecution narrative. In asylum evaluations, the evaluator elicits a detailed account of what the applicant experienced or fears: what happened, when, where, who committed the acts, what methods were used, what happened afterward, and why the applicant fears return. This is not a general psychological history — the assessment requires a specific, sequenced narrative that can be matched against country-condition evidence and assessed for internal consistency. The verbal account given in the clinical interview is the unfiltered version of that narrative, before it has been shaped into a formal written declaration through attorney preparation.

Psychological diagnoses tied to persecution. Immigration evaluations typically include a DSM-5 diagnosis — PTSD, major depressive disorder, anxiety disorders — and an explicit clinical nexus statement connecting the diagnosis to the claimed persecution. The evaluator probes symptom criteria, trauma history, onset and duration, functional impairment, and social context during the interview. That clinical exploration is the basis of the expert opinion that appears in the written report.

Country-condition specifics and persecutor identification. In asylum and related evaluations, the evaluator asks the client about conditions in the country of origin as experienced firsthand — who the persecutors were, whether they held official roles (police, military, government officials, paramilitary actors), whether the persecution targeted a family, community, ethnic group, or religious community. The clinical interview captures these specifics at a level of detail that identifies the country, the region, the organizations involved, and the individuals by name or role.

Crime disclosure for U-visa and T-visa evaluations. U-visa applicants must demonstrate substantial physical or mental abuse in connection with specific criminal activity — assault, domestic violence, sexual assault, trafficking. T-visa applicants must demonstrate trauma consistent with human trafficking. The clinical interview for these evaluations elicits a detailed account of the criminal conduct the applicant experienced, including method, perpetrator description, and circumstances. That verbatim crime disclosure is the foundation of the expert's opinion.

Immigration status woven throughout. A clinical interview for an immigration evaluation cannot be conducted without establishing the legal context — why the applicant is seeking this evaluation, what proceeding it is for, what their current immigration status is. The client's undocumented status, prior proceedings, entry history, and pending case details are embedded in the interview record from the first exchange.

What cloud AI scribes capture from an evaluation session

When a cloud AI scribe is running during the clinical interview for an immigration psychological evaluation, the vendor's data pipeline processes the complete session audio. As the cloud data-flow analysis documents, that pipeline typically involves raw audio upload to cloud infrastructure, intermediate transcript generation, note drafting via a cloud language model, and retention policies for each artifact tier under the vendor's data retention agreement.

From an immigration evaluation session, the vendor's retained data includes:

• The complete persecution narrative in the client's own words — the pre-declaration, unfiltered account of what happened, before any attorney preparation of the formal written declaration
• The evaluator's specific follow-up questions, which may have elicited additional detail about persecutors by name, locations, methods, and the client's specific fear of return
• The client's immigration status, entry history, prior proceedings, and current legal situation as described in the evaluation context
• Crime disclosures for U-visa or T-visa evaluations — the verbatim account of the criminal conduct the applicant experienced
• Country-of-origin specifics, including the region, locality, organizations, and individual actors identified by the client during clinical interview
• Any discussion between the evaluator and client about the attorney's legal theory, the specific issues the evaluation is addressing, or the pending immigration proceeding

This is not a record of a therapy session — it is a verbatim record of the most sensitive material the evaluation produces, held on vendor infrastructure as an independent artifact, separate from the evaluator's own written report and separately reachable by legal process.

Who can reach the vendor's evaluation data

USCIS in benefits adjudication. In asylum cases, USCIS adjudicates affirmative applications and can request records supporting the submission. An applicant submits the evaluator's written report as evidence; the clinical interview that underlies the report is the factual basis of the expert's opinion. A vendor independently holding session audio from that interview is a potential source of information about the applicant's stated grounds for protection, accessible to USCIS through legal process directed at the vendor as a separate entity.

DHS (ICE) in removal proceedings. Individuals in removal proceedings before immigration courts face DHS as an adversarial party. DHS has broad mechanisms to seek evidence relevant to removal proceedings. A cloud AI scribe vendor holding session audio from an immigration evaluation presented as evidence represents a potential independent record source — a verbatim account of the client's stated basis for seeking protection, accessible to government counsel independently of the evaluator's own files.

Immigration courts. Immigration judges have authority over evidence in removal proceedings. A subpoena or similar process directed at a cloud AI scribe vendor for session data from an evaluation admitted as evidence in proceedings before an immigration judge is a distinct legal process from any process directed at the evaluator or attorney.

Criminal proceedings. If the underlying facts involve criminal conduct — as they do in U-visa and T-visa cases — or if criminal allegations arise in connection with the immigration matter, a criminal subpoena directed at the vendor for session audio is within the range of legally available process. As the subpoena explainer documents, cloud vendor records are reached by legal process directed at the vendor as a separate custodian, independent of any privilege the evaluator or attorney holds over their own files.

Mutual Legal Assistance Treaty (MLAT) processes. When a foreign government initiates a criminal investigation of a person in the United States — a prosecutorial strategy sometimes used to undermine persecution claims by alleging that the asylum seeker committed crimes in the country of origin — the foreign government can request US government assistance in obtaining evidence through MLAT procedures. US-based cloud server records are within the potential scope of MLAT-driven legal process in ways that records physically existing only on an evaluator's device are not. This is a developing and unsettled area of law, but the structural exposure is real: vendor-retained audio containing a client's verbatim account of their persecution, their country of origin, and the actors involved exists on cloud infrastructure with a broader accessibility surface than records that never left the evaluator's office.

Attorney referral and work-product protection

Immigration psychological evaluations are commonly attorney-directed: an immigration attorney retains the evaluator, provides the case background, and submits the written report as evidence. This raises a natural question about whether attorney work-product doctrine insulates the evaluation from disclosure in adversarial proceedings.

Work-product doctrine protects documents prepared in anticipation of litigation at attorney direction. The evaluator's written report, prepared at attorney direction for submission in an immigration proceeding, may carry work-product protection in the hands of the attorney and evaluator. However, a cloud AI scribe vendor is not the attorney's agent. As the BAA explainer documents, a business associate agreement governs what a vendor can do with data it already holds — it does not transform the vendor's independently-retained session audio into an extension of the attorney's work product. The vendor's retained copy of the clinical interview is held by a third-party business associate under its own data retention policy. Any work-product argument against the vendor's production would need to be made independently against the vendor in the specific proceeding — a more complex undertaking than asserting privilege over the attorney's or evaluator's own files, and one that lacks the established doctrinal footing that applies directly to attorney-directed documents.

On-device processing and the single-custodian evaluation record

When the clinical interview for an immigration evaluation is recorded and processed entirely on the evaluator's device — transcribed locally via Whisper.cpp, drafted locally via an on-device language model on Apple Silicon, never transmitted to cloud infrastructure — the verbatim session audio has one custodian: the evaluator. The written report submitted to USCIS or the immigration court is the primary evidence document. The session audio and any working transcript are on the evaluator's device, subject to the evaluator's own legal obligations and privilege protections — not to a vendor's independent retention policy and not to the broader accessibility surface that cloud-stored records carry.

For evaluators who conduct multiple immigration assessments — forensic psychologists specializing in asylum work, immigration-focused LCSWs, and trauma-specialized LPCs accepting attorney referrals — on-device processing means no vendor holds a growing archive of their clients' persecution narratives, crime disclosures, country-of-origin specifics, and immigration status information across dozens of cases. That archive, if it existed on cloud infrastructure, would represent an extraordinarily sensitive collection of records for an extraordinarily vulnerable population: individuals whose legal protection depends precisely on the confidentiality of the claims they made in the evaluation session.

The mandated reporting analysis and the couples therapy discovery analysis both document how cloud vendor records become independently reachable in adversarial legal proceedings. For immigration evaluation clients, whose adversarial proceedings may involve the US government or foreign governments with MLAT access, the architectural guarantee that no cloud vendor holds that data is not a theoretical benefit — it is a concrete protection with real legal significance.

Further reading

• Can an AI therapy note be subpoenaed? A 2026 legal-risk explainer — the general mechanisms by which civil and criminal subpoenas reach cloud AI vendor records directly, including what clinicians cannot control about a vendor's response to process directed at the vendor
• The 7 things Mentalyc, Upheal, and Blueprint actually send to their servers — the request-by-request data flow behind cloud AI scribes: what artifact tiers the vendor retains and what each tier contains
• What is a BAA, actually — and what it does NOT cover — why a signed business associate agreement does not prevent a vendor's independently-held records from being compelled by legal process directed at the vendor as a separate entity
• Mandated reporting and AI scribes: what therapists document when a client discloses abuse — the parallel analysis for disclosure sessions: what the HIPAA mandatory-reporting exception does not cover for vendor-retained records and what categories of proceedings can reach that vendor data
• Couples therapy records, divorce discovery, and the AI scribe — how third-party vendor data custody complicates privilege assertions in adversarial legal proceedings involving clinical records

This post is educational commentary, not legal, clinical, regulatory, or immigration advice. Immigration law, HIPAA, state privilege frameworks, the court-order and subpoena provisions of HIPAA at 45 CFR 164.512(e), immigration court procedural rules under EOIR, and the evidentiary standards of USCIS adjudications are complex, fact-specific, and jurisdiction-dependent. Mutual Legal Assistance Treaty processes and their intersection with cloud data stored in the United States are a developing area of law. The application of attorney work-product doctrine to third-party vendor records is litigated case by case. Evaluators conducting immigration psychological assessments should consult a licensed healthcare attorney and a qualified immigration law professional before making documentation, technology, or record-retention decisions for their practice.