Russia is on the edge of a general artificial-intelligence law without yet having one. So the profile is a sophisticated system with real enforcement muscle and a genuine adjudicated AI docket, poised at the moment before its first general AI law takes effect.
The shape is a dense jurisdiction whose AI-specific binding law is one signature away. The soft tier holds an AI strategy and an ethics code; the binding tier is general law an AI dispute would meet, plus one lapsed AI experiment; and the single most consequential item sits in its own tier, adopted but not yet operative.
The core data-protection statute. A 2024 package sharpened its teeth: Federal Law No. 420-FZ (30 Nov 2024) rewrote the administrative-offences article to add tiered fines, with repeat leaks carrying a turnover fine of one to three per cent of annual revenue, and a companion criminal law added an offence for unlawful cross-border transfer. The statute most likely to govern an AI or data harm.
Source: pravo.gov.ruThe core information statute. Federal Law No. 408-FZ (31 Jul 2023) inserted Article 10.2-2, a recommender-technology transparency duty in force from 1 October 2023 that requires services using recommendation algorithms to inform users and publish the rules. A duty about algorithmic recommendation generally, not an AI-specific rule.
Source: pravo.gov.ruRequires the personal data of Russian citizens to be recorded and stored on servers located in Russia. General data-infrastructure law that shapes where AI training and processing may occur.
Source: pravo.gov.ruThe traffic-control and network-sovereignty regime for the Russian internet segment. General infrastructure law, not AI-specific.
Source: pravo.gov.ruThe general framework for regulatory sandboxes in digital innovation, including artificial intelligence. Enabling law that relaxes named rules for approved experiments; not a duty-imposing AI statute.
Source: pravo.gov.ruThe one earlier binding instrument written for AI: a five-year experiment in the city of Moscow that loosened data rules for AI development. Its term ran out in 2025; certain data mechanisms were carried forward by a 2024 amendment, but the experiment itself has lapsed.
Source: pravo.gov.ruThe intellectual-property code. Article 1347 confines the author of an invention to the natural person by whose creative work it was created; Articles 1257 and 1228 confine the author of a work to a natural person. The free-use provisions are a closed list in Articles 1273 to 1280. Neither addresses artificial intelligence.
Source: WIPO Lex; official Civil Code textThe state AI-policy framework, updated in 2024 to run to 2030, setting targets for adoption across industry and the public sector. A planning act that tasks state bodies and creates no private-operator duties.
Source: pravo.gov.ruA voluntary ethics code signed by the major technology companies, with an ethics commission from 2022. It contains content-marking and human-oversight norms, but as a signed corporate code it binds only its signatories by their own undertaking, not by law.
Source: a-ai.ruRussia's first general AI statute. As reported it defines sovereign and national foundation models, requires Russian hosting and alignment with traditional values, and adds a training and copyright carve-out. Its content-marking provision requires large platforms, those with 500,000 or more daily users, to provide the means to label AI-generated content, though the labeling itself is not mandatory. General effect is set for 1 September 2026 and the labeling provision for 1 March 2027, so nothing binds yet.
Source: State Duma and Federation Council records; Russian legal pressConfirmed founding member; signed in Shanghai on 16 Jul 2026; Russian signatory Maxim Oreshkin, Deputy Chief of Staff of the Presidential Executive Office. Official source: China State Council release; Chinese foreign-ministry readout of the signing.
Adopted 6 Jul 2025 at the Rio summit under the Brazilian presidency; Russia a member. Official source: BRICS presidency documents.
Summit-level political commitments
Adopted 29 May 2026 in Astana by the Supreme Eurasian Economic Council; Russia a member. Official source: Eurasian Economic Commission.
Adopted by the CIS Inter-Parliamentary Assembly on 18 Apr 2025; a non-binding template; Russia a member. Official source: Inter-Parliamentary Assembly of the CIS.
Adopted by consensus in 2024; Russia bound as a UN member; no co-sponsorship asserted. Official source: UN records.
Adopted by consensus 22 Sep 2024; Russia bound as a UN member, having led a dissenting amendment that a no-action motion set aside. Official source: A/RES/79/1 annex; UN records.
Adopted 2021, as a UNESCO member; no readiness assessment found. Official source: unesco.org.
Axes: specificity (x, 0-100) is how squarely the instrument is written for AI; force (y, 0-100) is how hard it binds. Placements are editorial judgments for the comparison plot; the tier column is the authoritative classification. The pattern is a country whose binding tier is general and forceful but not AI-specific, whose AI-specific texts are either soft or adopted-but-not-yet-operative, and whose regional band is unusually full.
Tap a marker
The binding and prescriptive corner is the crowded one: the CAC-led measures put force and detail together, service by service. The sector guidance sits just left of the binding line, nonbinding in form but treated as obligatory. Dashed markers are drafts; the National AI Law would consolidate the whole stack into one statute.
Russia has a genuinely adjudicated AI matter, unusual among newer profiles: a deepfake copyright case that ran through three instances with a citable docket number. Around it sit reported matters, a facial-recognition challenge with an outcome but no citable number, a deepfake-fraud pattern, and a national case review. Per the fabrication screen, only the one verified number is asserted; everything else is recorded without a number.
A single-source account of an AI-generated confession video refused as evidence could not be verified and is not treated as a documented matter.
Russia holds the human line in the Civil Code while its institutions move: authorship and inventorship stay with natural persons and the closed free-use list has no mining lane, but a framework AI law awaiting signature carries a training carve-out, and the patent office already deploys AI in its own examination.
The AI-IP Index reads a country on two questions that usually get folded together. Accommodation asks how far the law will go to protect intellectual property that a machine helped produce. Institution asks how much administrative machinery sits behind that law: its guidance, its international submissions, its disclosure duties.
A country can be generous on paper and have built almost nothing, or cautious and highly organised, so the two are scored apart and reported as a pair. Each rests on three questions, scored 0, 0.5 or 1.
How far the law will protect intellectual property that a machine helped make.
How much administrative machinery sits behind that law.
The scale runs the same way on both axes: 0 is the open or developed end, 1 is the restrictive or absent one, and 0.5 sits between for anything conditional or untested. Read as a pair, the two scores show whether AI-assisted work can be protected here, and whether the office has said so in writing.
Within each axis the three questions carry equal weight. That is a deliberate choice, not an oversight. An AI developer might reasonably value freedom to train above patent inventorship, but weighting is where indices lose their credibility, so the score stays a flat mean and the choice is stated in the open. Each cell also cites a statute section, a case or an office page, so any single score can be challenged without disputing the rest.
Two positions are worth naming. A country can be open on the law but thin on machinery, an open-but-undeveloped stance that bets on ambiguity. Or it can be restrictive on rights yet run an active office, a restrictive-but-developed stance that reflects a settled policy choice rather than a gap. The pair of scores tells those two apart where a single ranking would blur them.
A rubric is only as good as the results it produces on cases where the answer is already known. These three jurisdictions sit outside this series and are scored with the same six questions, as a check. Their law is the best documented anywhere, so if the scale puts them where the law actually sits, the scale is doing its job. They are reference points, not editions.
Firm judicial settlement at both poles and the weakest statutory position on training. The courts have closed the door on AI inventors and AI authors, while everything on training rides on unresolved fair-use litigation. There is no per se duty to declare AI use in a filing, only a candour duty that bites where the use is material.
Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022); Thaler v. Perlmutter (D.C. Cir. 2025), cert. denied 2 Mar 2026; Bartz v. Anthropic (N.D. Cal. 2025); USPTO revised inventorship guidance, 90 Fed. Reg. 54636 (28 Nov 2025); 89 Fed. Reg. 58128 (17 Jul 2024); 89 Fed. Reg. 25609 (11 Apr 2024).
The most permissive statutory position on training and the strongest transparency duty in force, sitting over an authorship question no court has answered. Read the label carefully: training and transparency run through EU law, but patents run through the EPO under the European Patent Convention, which is not an EU institution and covers 39 states.
Directive (EU) 2019/790, arts. 3 and 4; Regulation (EU) 2024/1689, art. 53(1)(c) and (d); EPO J 8/20 (2021); EPO Guidelines G-II 3.3.1 (2026 edition); Infopaq C-5/08; Painer C-145/10.
Liberal on patents since the Supreme Court abandoned the Aerotel test in February 2026, conservative on copyright, and currently thin on AI-specific machinery after the dedicated AI examination guidance was withdrawn. Section 9(3) nominally protects a work with no human author for 50 years, the widest AI-output right on paper, but no court has ever applied it to generative AI.
CDPA 1988 ss. 9(3), 12(7), 29A, 178; Thaler v Comptroller-General [2023] UKSC 49; Comptroller-General v Emotional Perception AI [2026] UKSC 3; UKIPO practice notice, 14 Jul 2026; Data (Use and Access) Act 2025 ss. 135 to 137.
The check earned its keep twice. The untested rule held: section 9(3) of the UK Copyright Act reads like the widest AI-output right anywhere, yet because no court has applied it to generative AI it scores as untested rather than open, which is the correct answer and the one a headline reading would miss. The check also exposed a limit worth stating. The Institution axis measures AI-specific machinery, not general office capacity, so a long-established office can score poorly when it has withdrawn its AI guidance and imposes no disclosure duty. Read that axis as what an office has built for AI, nothing wider.
Each row below carries the test we applied and the finding behind the score, with a source. The rationale, not the number, is the point.
Scores verified 21 July 2026, against AIP Index codebook v1.0. A score is only true as of its date.
What we look forA granted patent or a clearly workable examination route, not merely a silent statute.
The findingThe Civil Code confines the author of an invention to the natural person by whose creative work it was created, so an AI-assisted invention with a human inventor is open in principle while the machine cannot be named. No AI-specific examination practice, guidance or grant evidence exists, and no DABUS-type filing has produced a Russian grant.
What we look forProtection applied where human choices shape the work, and refusal of purely machine output.
The findingThe Civil Code confines the author of a work to the natural person whose creative labour created it, and excludes those who give only technical or organizational support, so a human who directs an AI tool and contributes authorship can hold copyright while purely AI-generated output has no qualifying author. There is no computer-generated-works provision; a state bank's 2021 registration was a human-authored software registration, not an AI-authorship grant, and the point is otherwise untested.
What we look forA statutory or judicial mining lane. Silence scores against the miner, since the reserved-rights default governs.
The findingThere is no training or text-and-data-mining exception in force. The free-use provisions are a closed enumerated list in Articles 1273 to 1280 with no mining carve-out, and the terms text-and-data mining do not appear in the code. With copyright silent, the reserved-rights default governs the use of works for training. The pending AI law would add a training carve-out; that change is recorded apart, under trajectory, because it is not yet in force.
What we look forA traceable, substantive submission on the record, not bare membership.
The findingRussia is a major WIPO member and its patent office is historically active in international fora, but no traceable submission from Russia or Rospatent to the WIPO Conversation on intellectual property and AI could be verified, and post-2022 isolation is a plausible factor. Membership and activity without a traceable submission earns the middle.
What we look forA standalone AI examination text, not AI handled quietly under general practice.
The findingRospatent has no published dedicated guidance on examining AI-related filings, but it deploys AI in its own examination search and its leadership has publicly signalled readiness to register AI-assisted inventions. That is partial institutional engagement rather than none, so it sits at the middle rather than the open pole, and is the reason the institution axis reads below the thin-jurisdiction level.
What we look forA mandatory declaration inside the filing, not a proposal or an informal request.
The findingNo in-force instrument requires declaring AI use in a filing or imposes a labeling or transparency duty on AI-generated content. The framework AI law's labeling and disclosure provisions are not in force and are not filing-disclosure duties in any event.
Method. Six questions, three per axis, scored 0, 0.5 or 1 against AIP Index codebook v1.0. The three questions in each axis are weighted equally by choice, and the two axes are reported as a pair, never blended into one number. Every score cites a statute, case or office page, carries a verified date, and records its direction of travel apart from the level. Positions are editorial judgements read from primary law and office practice, not official scores. WIPO's public catalogue of IP-office AI initiatives does not score jurisdictions on this basis, so the readings here come straight from the source law. Not legal advice.
Russia is on the edge of a general artificial-intelligence law without yet having one. In July 2026 the Federal Assembly adopted a framework AI statute: the State Duma passed it on 7 and 8 July and the Federation Council approved it on 17 July. As of this dossier it is not yet in force. It awaits the President's signature, no federal-law number has been published, and its own text defers general effect to 1 September 2026 and its content-labeling provision to 1 March 2027. Until then Russia governs AI through instruments written for other purposes. A national AI strategy and a corporate ethics code sit in the soft tier. The binding tier is dense but general: a much-amended data-protection law now backed by turnover-scaled fines, an information law that added a recommender-algorithm transparency duty, data-localization and network-sovereignty statutes, a regulatory-sandbox framework, and the intellectual-property code, which confines authorship and inventorship to natural persons. The one earlier binding instrument written for AI, the Moscow experimental legal regime, was a five-year experiment that lapsed in 2025. So the profile is a sophisticated system with real enforcement muscle and a genuine adjudicated AI docket, poised at the moment before its first general AI law takes effect.
The site does three things: it maps every instrument that governs AI in Russia, 18 of them, ordered by how hard each one binds; it reads the same instruments a second time by how squarely each is written for AI; and it records the matters where AI meets the law, with what was searched and found empty stated as plainly as what was found. Every claim traces to a primary source, and absences are recorded rather than papered over.
A membership claim enters this table only from an official list: a declaration annex, a treaty signature table, an organization's own record or an official government statement. Russia's regional band is the fullest in the series, and its Global Digital Compact line carries the distinctive nuance that it is bound by consensus while having formally dissented; the WAICO entry is set out below.
Know of a ruling or proceeding in Russia where AI is central? Send a source link; it will be reviewed before publishing.
Submit a case