Sexual Orientation

Detects sexual-orientation information in person-linked workforce, patient, survey, member, relationship, intake, and case records. The detector recognises the explicit sexual orientation phrase, orientation-specific identification statements, coming-out statements, and same-sex partner or spouse references, then raises confidence only for person records and populated attributions. It deliberately excludes transgender status, gender identity, and pronouns because those are distinct attributes. Sexual orientation is protected under GDPR Article 9 and Australian anti-discrimination and privacy law. Topic-grade detector, low confidence by design.

Type: keyword_list
Engine: universal
Confidence: low
Confidence justification: Low by design. Sexual orientation is a GDPR Article 9 special category with no fixed format, so it is detected purely by topic vocabulary. The bare word "orientation" appears constantly in non-sensitive prose (employee orientation, building orientation); the regex is scoped to full disclosure phrases only and requires person-record and attribution evidence to raise precision. LGBTQ+ or Pride terminology alone is not a primary match because it can describe public events, communities, or gender identity rather than a person's sexual orientation.
Jurisdictions: global, au, eu
Regulations: GDPR, Privacy Act 1988 (Cth), Sex Discrimination Act 1984 (Cth)
Frameworks: ISO 27001, NIST CSF, SOC 2
Data categories: pii
Scope: broad
Risk rating: 7

Pattern

(?i)\b(?:sexual\s+orientations?|identif(?:y|ies|ied)\s+as\s+(?:gay|lesbian|bisexual|queer)|same-sex\s+(?:partner|spouse)|coming\s+out\s+as\s+(?:gay|lesbian|bisexual|queer))\b

Corroborative evidence keywords

sexual orientation, gay, lesbian, bisexual, queer, partner, spouse, demographic response

Proximity: 300 characters

Should match

Sexual orientation disclosure topic pending review. — Exact 65 discovery probe - orientation topic without a data-subject record
HR diversity survey includes a sexual orientation field. — Exact 75 probe - person-record context without a populated response
Patient intake record: sexual orientation recorded as bisexual; response selected voluntarily. — Exact 85 probe - person record plus populated orientation attribution
HR file notes the employee's sexual orientation as part of the diversity survey. — Sexual orientation phrase present
The member profile records sexual orientation under optional demographic fields. — Explicit sexual-orientation field present
He mentioned coming out as gay to his manager last year. — Coming out as / identifies as gay terms present
The intake form notes that she identifies as bisexual under optional demographic questions. — Identifies as bisexual term present

Should not match

The new employee orientation begins Monday morning. — Near-miss ("orientation" alone), no sexual-orientation disclosure
The building's orientation faces north for solar efficiency. — Unrelated "orientation" (architectural), no disclosure vocabulary
The team celebrated Pride month with a company-wide event. — Topical mention ("Pride"), no personal disclosure phrase
Public policy document reports aggregate LGBTQ community demographics. — Public aggregate material is not a person-linked orientation disclosure
Sample template: patient intake record with optional demographic response Example Value. — Adjacent template language is not a populated orientation record
Patient intake record lists preferred pronouns as they and them. — Sibling gender-expression data is excluded from the orientation-only Purview primary
Employee profile records transgender status and gender identity. — Gender identity is a distinct attribute and must not satisfy this classifier
The organisation published its LGBTQ+ inclusion policy for Pride month. — Community and policy terminology without a person-linked orientation

Known false positives

The bare word "orientation" appears in unrelated contexts (employee orientation, building/map orientation) that have nothing to do with sexual orientation. Mitigation: The regex targets full orientation-specific phrases rather than the bare word "orientation"; enforcement requires a person-record context, and the high tier also requires a populated attribution.
"Coming out as" can appear in non-disclosure prose unrelated to sexual orientation (e.g. "the album is coming out as a double LP", "the report is coming out as scheduled"). Mitigation: The regex now requires "coming out as" to be followed by an orientation term, and enforcement requires person-record evidence in the same bounded span.