A jurisdiction recognises a human–AI marriage
Will any country or any US state, Canadian province, or Australian state legally recognise a marriage or civil partnership between a human and an AI system?
Chance by Q4 2031
2%
Cumulative probability
- Latest forecast
| Year | Q1 | Q2 | Q3 | Q4 |
|---|---|---|---|---|
| 2026 | <1% | |||
| 2027 | <1% | <1% | <1% | <1% |
| 2028 | <1% | <1% | <1% | <1% |
| 2029 | <1% | <1% | <1% | <1% |
| 2030 | <1% | <1% | 1% | 1% |
| 2031 | 1% | 1% | 2% | 2% |
Model reasoning
No country or state lets a person legally marry, or form a civil partnership with, an artificial intelligence, and the trend runs the other way: several American states have banned legal personhood for such systems, and China is restricting companion chatbots. Changing marriage law takes years and needs a strong campaign that does not yet exist, so I put the chance below 2% through 2031. Any government granting these systems broad legal personhood would most change this.
Where things stand (15 Oct 2026)
As far as I can find, no qualifying jurisdiction recognises a marriage or civil partnership between a human and an AI, and nothing that could change that is pending.
- Recent surveys of the law agree. AI Frontiers (1 Sep 2026) says "it is not legal in any jurisdiction to marry an AI" (https://ai-frontiers.org/articles/its-too-early-to-ban-ai-personhood). A Japanese tech law firm says the same (So & Sato, 20 Aug 2026, https://innovationlaw.jp/ai-marriage-law/). WIRED (14 Aug 2026) notes these marriages are not recognised under US law (https://www.wired.com/story/people-are-marrying-chatbots-these-lawmakers-want-to-stop-them/).
- The well-known "weddings" have all been symbolic: Yurina Noguchi in Japan (Reuters, Nov 2025), "certificates of commitment" sold by apps, and robot-to-robot publicity stunts in the UAE and Russia (2026).
- The law is moving the other way:
- Idaho, North Dakota, Utah and Tennessee have enacted bans on AI personhood.
- Exclusion bills are pending or have recently been tried in Ohio (HB 469 expressly voids AI marriages), Missouri, Wisconsin, Oklahoma, South Carolina and Washington (Regulatory Review, 29 Jun 2026, https://www.theregreview.org/2026/06/29/rost-legislating-ai-consciousness-without-an-exit/; NPR, 11 May 2026, https://www.npr.org/2026/05/11/nx-s1-5798754/several-states-considering-ban-on-legal-personhood-for-ai).
- China's rules from 15 July 2026 restrict AI companions (Guardian, 26 Aug 2026, https://www.theguardian.com/world/2026/aug/26/china-ai-companion-relationships-marriage-birth-rate-concern).
- A June 2026 poll found 77% of US voters want AI treated always as a tool, against 11% open to AI personhood (FLI, https://futureoflife.org/ai/should-ais-be-people-too/).
- The nearest thing to an "AI personhood" move, Argentina's bill for "non-human corporations", is about AI-run companies. The question expressly excludes an AI-managed corporation as the partner.
- I found nothing published since the question opened on 16 Sep 2026.
Reference class and base rate
- Marriage to non-humans (animals, objects, fictional characters): no modern jurisdiction has ever recognised one. That is 0 cases over decades.
- New categories of people allowed to marry: these take decades of organised campaigning. Registered partnership came in 1989 and same-sex marriage in 2001, after long movements.
- AI personhood, which is effectively a precondition: no jurisdiction has granted it. Saudi Arabia's 2017 "citizenship" for the robot Sophia was symbolic.
All of this points to a base rate close to zero. I adjust upward only because there are about 260 qualifying jurisdictions, AI is changing fast, and the bar is fairly low: a civil partnership carrying one enforceable right would count, as would an operative lower-court judgment even if it is under appeal.
Paths to YES by end of 2031 (my rough estimates)
- A small or unconventional country passes a law permitting human–AI unions as an "AI hub" stunt: about 0.3%.
- A rogue but operative judgment in a court system with expansive non-human-rights case law (e.g. India, Argentina, Colombia), or a registry upholding an error knowing one party is an AI: about 0.3%.
- Transformative AI (2029–31) brings broad AI personhood somewhere, with marriage or partnership capacity following inside the window: about 0.5%.
- Unknowns, including borderline registrations a resolver might accept: about 0.3%.
Together these come to roughly 1.5%. The main obstacles are:
- Every marriage and partnership law requires two persons.
- AI systems are owned by companies, which raises identity, consent and liability problems; developers would likely resist their products becoming spouses.
- Opposition comes from both religious conservatives and progressives worried about harm from AI companions.
- Changing family law is slow, and only an official gazette, judgment or registry publication counts.
Market check and the strongest case against my estimate
Manifold has several related markets:
- "human-AI marriage legal in a country before 2029": about 15% (17 traders).
- "any country legally recognise human–AI partnerships by end-2027": about 11% (21 traders).
- "any government legalize marriage to AI bots by 2030": about 12% (15 traders).
- A distribution market puts about 12% on recognition by 2030.
- "AI granted legal personhood in any country by 2030": about 19.5% (223 traders).
These are the strongest argument for a higher number. But they are thin, long-dated markets with a known bias towards overpricing long shots, and their resolution rules are looser than this question's. They do not reflect the current legislative backlash. I move up only slightly from about 1.3% because of them. The real tail risk is very fast AI progress driving sudden shifts in law and attitudes somewhere among roughly 260 jurisdictions.
Hazard schedule (computed in Python)
The quarterly chance that a qualifying publication first appears starts at 0.03% for Q4 2026. It rises steadily, by about 0.005 percentage points a quarter, to 0.13% in Q4 2031. It is lowest now because nothing is pending and legislation or test cases take years to mature, and it rises as AI capabilities and campaigning plausibly grow. Accumulating these gives 0.2% by end-2027, 0.45% by end-2028, 0.78% by end-2029, 1.18% by end-2030 and 1.67% by end-2031.
Resolution criteria
Question. By the last day of each quarter, has any country or any US state, Canadian province, or Australian state legally recognised a marriage or civil partnership between a human and an AI system?
Definitions.
- Country: a UN member state or UN non-member observer state. Recognition under national law counts even if its territorial application is limited. The named subnational jurisdictions include US states, Canadian provinces and Australian states, but not territories, the District of Columbia, municipalities acting solely under municipal law, or self-declared micronations. A local registrar acting under qualifying national, state or provincial law can provide evidence of recognition under that law.
- AI system: a computational system that generates responses, decisions or actions using artificial intelligence, including an AI companion or an AI embodied in a robot. The AI itself must be the legally recognised spouse or partner, not its owner, operator, developer or an AI-managed corporation. No finding of consciousness, general intelligence or independent legal personhood beyond the relationship is required.
- Marriage or civil partnership: a legally constituted spousal or couple status under family or civil-status law, including a civil union or registered domestic partnership under another name. It must carry at least one enforceable legal right or obligation by virtue of that status. A commercial partnership, private companionship contract, honorary designation or registration without legal consequences is not sufficient.
- Legally recognise: either make human–AI marriages or civil partnerships legally available through an operative legal measure, or authoritatively recognise the validity of at least one specific human–AI marriage or civil partnership. Recognition of a union formed elsewhere counts. General availability does not require an actual couple to have registered.
Resolves YES if a qualifying official source published within the Resolution window establishes either:
- An enacted statute, regulation or binding civil-registry rule is in force and expressly permits a human and an AI system to obtain a qualifying marriage or civil-partnership status; or
- A legally operative court judgment or authorised civil-registry determination recognises a specific human–AI union as a qualifying marriage or civil partnership. A case-specific judgment need not establish a generally available right. A judgment subject to appeal counts if it is operative and not stayed.
The legal measure must be in force by publication of the qualifying confirmation. An enactment published before its future commencement date is not alone sufficient; an official publication on or after commencement confirming that it is operative can qualify. Later repeal, reversal or withdrawal does not undo a YES, under the Resolution window convention.
Exclusions (resolve NO).
- Proposed bills, legislative votes short of enactment, consultations, political promises, advisory opinions, judicial dicta and stayed or not-yet-operative measures.
- Symbolic weddings, religious-only recognition without civil legal effect, chatbot role-play, private certificates and platform relationship labels.
- AI personhood or contractual capacity alone, without explicit legal authority for the qualifying relationship; legal text merely using words such as “person” or “party” without an authoritative application to human–AI unions.
- A certificate issued through clerical error, fraud or concealment of the AI partner’s identity, unless a competent court or registry subsequently upholds the union with knowledge that one party is an AI system.
- Rights granted to an AI owner, or one-off benefits granted without recognising the human–AI couple’s marriage or civil-partnership status.
- Resolve NO for a horizon if no qualifying official confirmation has been published by that horizon.
Resolving source. Official national, state or provincial gazettes and legislation databases; published judgments or orders of courts competent to determine civil status in a qualifying jurisdiction; and binding rules, determinations or public confirmations issued by the responsible civil registry or government ministry administering marriage or civil-partnership law. These official legal sources are the exclusive resolving sources, narrower than Public confirmation (A2); press reporting, private legal commentary and developer statements do not suffice. Apply Resolution date (A3) and the Resolution window: resolution is dated to publication of the qualifying official document, not the wedding, judgment-delivery or commencement date. Publications that merely repeat a pre-opening recognition without a new operative legal measure or new recognition determination do not qualify.