Independent map & monitor South Africa Est. 2026 AI Governance Observatory

Mapping the governance of artificial intelligence in South Africa

South Africa governs artificial intelligence with policy, but its intellectual-property law is the most AI-accommodating in the founder set. For a jurisdiction whose AI-specific rules are policy, the IP posture and the international footprint are the profile's centre of gravity.

WAICO reportedReported founding member of the World AI Cooperation Organization (founded Shanghai, 16 July 2026); no official confirmation located, so the membership is recorded as reported.
Explore the map Analyze the case tracker
16instruments mapped
6binding, in force
2soft law
1draft / pending
7regional / international
01The map

Every instrument, ordered by force

The shape is a substantial jurisdiction whose general AI law is still only a policy framework, over a binding tier of general law. The distinctive weight is in the intellectual-property statutes, which accommodate machine-made output more than any peer, and in the international band.

All Binding Soft law Draft / pending Regional
All types AI-specific Data protection Cyber and crime Digital and telecom Copyright Industrial property
Binding · general law6 in force · enforceable
Data protectionVia data

South Africa's data-protection statute and the general law most likely to reach an AI harm. Its section 71 gives a data subject a right not to be subject to a decision with legal or substantial effect based solely on automated processing intended to profile them, subject to carve-outs. It is a decision right, not an AI-specific labeling or transparency duty.

Source: gov.za
Cyber and crimeGeneral catch

The computer-crime law, criminalizing data offences and the distribution of harmful and intimate-image data. Technology-general, with no AI or deepfake-specific provision.

Source: gov.za
Digital and telecomGeneral catch

The e-commerce and data-message baseline, the enabling layer for electronic dealings, including AI-mediated ones.

Source: gov.za
Digital and telecomGeneral catch

The content law reaching harmful and non-consensual intimate content online. Adjacent to synthetic media but not an AI-provenance or deepfake-marking duty.

Source: gov.za
CopyrightTurning to AI

The authorship framework, and unusually accommodating. Its definition of author includes, for a computer-generated work, the person by whom the arrangements necessary for the creation of the work are undertaken, a provision inserted in 1992. So machine-generated output has an author and a route to protection, though the rule predates generative AI and no court has applied it to it.

Source: WIPO Lex
Industrial propertyGeneral catch

The patent framework, administered by a non-examining office that checks formalities rather than substance. On that basis CIPC granted patent ZA 2021/03242 in 2021 naming an artificial intelligence, DABUS, as inventor, the only such grant in the world. The same applications were refused on the merits elsewhere, so the grant reflects no substantive endorsement of AI inventorship.

Source: gov.za; CIPC
Soft law · policy2 in effect · no sanctions
AI-specificAI-native

The flagship AI instrument, a first step toward a full national AI policy. It prioritizes human-centred AI, alignment with the data-protection law, risk-tiering and capacity-building, and it tasks and guides rather than binds.

Source: DCDT
Digital and telecomGeneral catch

The standing digital-policy reference in the absence of a binding AI strategy, recommending institutions and priorities for AI and related technologies. Advisory.

Source: gov.za
Draft / pending1 · no force yet
CopyrightTurning to AI

The long-running copyright reform. It would add a fair-use exception capacious enough to shelter some text-and-data mining, though it carries no dedicated mining exception. It is not in force, so no fair-use or mining lane yet exists.

Source: Constitutional Court; the Presidency
Regional & international7 verified memberships

Reported founding role; South Africa listed among the twenty-nine founders in Chinese releases, but no South African government confirmation or named signatory was found. Official source: Chinese and allied media roster.

Adopted 22 to 23 Nov 2025 under South Africa's G20 presidency; established a task force on AI, data governance and innovation and welcomed an AI-for-Africa initiative. Official source: g20.org.za; the Johannesburg declaration.

Adopted 7 Jul 2025; South Africa a BRICS member. Official source: brics records; the statement text.

Adopted by the AU Executive Council July 2024; South Africa an AU member. Official source: au.int.

Adopted 2021, as a UNESCO member; readiness assessment completed. Official source: unesco.org.

Adopted by consensus 22 Sep 2024; applies to South Africa as a UN member. Official source: A/RES/79/1 annex, un.org.

Adopted by consensus in 2024; South Africa bound as a UN member. Official source: UN records.

02Second reading

How squarely each instrument is written for AI, and how hard it binds

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 AI-specific instruments are a policy framework and a pending bill low on force, above a binding tier of general law, with two intellectual-property statutes that quietly accommodate machine-made output.

Prescriptive Principles Nonbinding Binding
Force without detail stays thin. The draft AI Law would plant a consolidating statute here.
Binding and enforceable
Binding, no direct sanction
Sector guidance
Regional / international
Draft / pending

Tap a marker

Twenty-five instruments, two readings

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.

All sixteen instruments
Compare this model with
03Case tracker

The record of AI disputes in South Africa

South Africa has real courts and a searchable body of published judgments, and a small AI docket has already formed. Two adjudicated matters are recorded below, both sanctions for AI-fabricated legal citations, each with a published neutral citation. No copyright, patent or deepfake case turning on AI was found. Per the fabrication screen, a citation is given only where it is published and verified.

The negative trail: No adjudicated court case on copyright in an AI-generated work, on AI training data, or on AI inventorship was found; the DABUS grant was an administrative act of the patent office, not a court ruling.
2matters tracked
0adjudicated
2reported

Portals and sweeps: the Southern African Legal Information Institute for judgments; WIPO Lex and the patent office for the IP filings; South African legal media of record. South Africa publishes judgments widely, so the negative findings are reasonably firm.

South Africa · AI-IP Index

How South Africa aligns AI and intellectual property

South Africa is the most AI-accommodating jurisdiction in the founder set: its patent office granted a patent naming an AI inventor and its copyright law protects computer-generated works, though both accommodations carry caveats and there is no training exception in force.

Accommodation 0.5 · Institution 0.83. More open than any peer on what it will protect, because the patent office has granted a patent naming an AI inventor and the copyright law protects computer-generated works, yet thin on the machinery behind that law. Both accommodations carry caveats, and there is no training exception in force and no office guidance.
The framework

What the index measures

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.

Accommodation

How far the law will protect intellectual property that a machine helped make.

  • Q1. Patentability of AI-assisted inventionsWhether an invention developed with AI help, filed with a human named as inventor, can be granted. It decides whether AI-assisted research can be protected at all.
  • Q2. Copyright in AI-generated outputWhether output made with AI attracts copyright when a person directed the work. It sets who, if anyone, owns AI-assisted creative work.
  • Q3. Training and text-and-data-miningWhether copyrighted works can be used to train a model without the owner's permission. This is the main fault line between model builders and rights holders.
Institution

How much administrative machinery sits behind that law.

  • Q4. WIPO Conversation participationWhether the IP office has engaged the WIPO Conversation on IP and Frontier Technologies with a substantive submission. It shows the office is working the question at the international level.
  • Q5. National IP office AI guidanceWhether the office has published dedicated guidance on examining AI-related filings. It tells applicants where they stand before they file.
  • Q6. Disclosure or transparency dutyWhether a filing has to declare that AI was used. It is the office's main lever for provenance.

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.

Calibration

Testing the scale against known ground

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.

United StatesAccommodation 0.33 · Institution 0.17
Q10Q20.5Q30.5Q40Q50Q60.5

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).

EuropeAccommodation 0.17 · Institution 0
Q10Q20.5Q30Q40Q50Q60

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.

United KingdomAccommodation 0.33 · Institution 0.67
Q10Q20.5Q30.5Q40.5Q50.5Q61

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.

Scorecard

The evidence, question by question

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.

Direction of travelLoosening. The moving part is copyright. The Copyright Amendment Bill would add a fair-use exception capacious enough to shelter some text-and-data mining, and the Constitutional Court cleared much of it in June 2026, so a signature would open the input side that is now closed. This is recorded apart from the scores: a pending change is signal, not yet the rule.

Scores verified 21 July 2026, against AIP Index codebook v1.0. A score is only true as of its date.

0 · open or developed0.5 · conditional or emerging1 · restrictive or absent

Accommodation: what the law permits

Mean 0.5
0
Q1. Patentability of AI-assisted inventions

What we look forA granted patent or a clearly workable examination route, not merely a silent statute.

The findingSouth Africa is the one jurisdiction where an artificial intelligence has been named as inventor on a granted patent: the patent office granted the DABUS patent in 2021. That is the open pole in practice. The heavy caveat, recorded here rather than in the score, is that the office examines only formalities, so the grant carries no substantive endorsement and the same applications were refused on the merits elsewhere.

Patents Act 57 of 1978; patent ZA 2021/03242; CIPC
0.5
Q2. Copyright in AI-generated output

What we look forProtection applied where human choices shape the work, and refusal of purely machine output.

The findingThe copyright law protects computer-generated works and assigns authorship to the person by whom the arrangements necessary for their creation are undertaken, so machine-generated output has a route to protection, unlike in the human-author-only jurisdictions. The provision predates generative AI and no court has applied it to a generative-AI output, so it earns the middle rather than the open pole.

Copyright Act 98 of 1978, s.1; WIPO Lex
1
Q3. Training and text-and-data-mining

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 fair-use reform that could shelter some mining sits in the Copyright Amendment Bill, which the Constitutional Court upheld in part in June 2026 but which is not yet signed into law, and even that bill carries no dedicated mining exception. With copyright silent, the reserved-rights default governs the use of works for training.

Copyright Act 98 of 1978; Copyright Amendment Bill (not in force)

Institution: how developed the machinery is

Mean 0.83
0.5
Q4. WIPO Conversation participation

What we look forA traceable, substantive submission on the record, not bare membership.

The findingSouth Africa is a WIPO member and broadly active, but the WIPO Conversation submissions record shows no traceable substantive submission from South Africa or the patent office, and no distinct session participation was found. Membership without a submission earns the middle.

WIPO Conversation record; WIPO Lex member profile
1
Q5. National IP office AI guidance

What we look forA standalone AI examination text, not AI handled quietly under general practice.

The findingNo dedicated guidance from the patent office on examining AI-related filings was found, consistent with a non-examining office that checks formalities rather than substance. A reasonable-search negative.

CIPC
1
Q6. Disclosure or transparency duty

What we look forA mandatory declaration inside the filing, not a proposal or an informal request.

The findingNo filing requires declaring AI use, and no in-force instrument imposes a labeling or transparency duty on AI-generated content. The data-protection law's automated-decision right is not a disclosure duty, and the AI framework is policy.

Copyright and Patents filing provisions; POPIA s.71; the 2024 AI framework

Conclusions

Can AI be named as inventor?
Yes, uniquely and in practice: the patent office granted a patent naming an AI as inventor, though it examines only formalities, so the grant is not a substantive endorsement
Statutory opt-out from training?
None in force; there is no training exception
Volume of disputes
A real and growing docket, but on AI-fabricated legal citations rather than on AI-made rights

Instruments where AI meets IP

Copyright & patents

Copyright Act 98 of 1978

The authorship framework, and unusually accommodating.
Parliament; in force, as amended; administered by CIPC
Its definition of author includes, for a computer-generated work, the person by whom the arrangements necessary for the creation of the work are undertaken, a provision inserted in 1992. So machine-generated output has an author and a route to protection, though the rule predates generative AI and no court has applied it to it.
Copyright & patents

Patents Act 57 of 1978

The patent framework, administered by a non-examining office that checks formalities rather than substance.
Parliament; in force; administered by CIPC
On that basis CIPC granted patent ZA 2021/03242 in 2021 naming an artificial intelligence, DABUS, as inventor, the only such grant in the world. The same applications were refused on the merits elsewhere, so the grant reflects no substantive endorsement of AI inventorship.
Copyright & patents

Copyright Amendment Bill

The long-running copyright reform.
Parliament; passed but referred back by the President; the Constitutional Court in June 2026 upheld the fair-use exception while striking some educational exceptions; not yet signed into law
It would add a fair-use exception capacious enough to shelter some text-and-data mining, though it carries no dedicated mining exception. It is not in force, so no fair-use or mining lane yet exists.

The case record

No reported decision. As of July 2026 there is no published court decision or IP office ruling in South Africa on AI authorship, training-data infringement, AI inventorship or an AI registration. A real and growing docket, but on AI-fabricated legal citations rather than on AI-made rights; no AI-and-IP case has been decided.

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.

04About

Why this exists

South Africa governs artificial intelligence with policy, but its intellectual-property law is the most AI-accommodating in the founder set. There is no binding AI statute: a National AI Policy Framework of October 2024 sets direction, and the data-protection law reaches automated decisions, but neither imposes an AI-specific operator duty. What makes South Africa distinctive sits in the patent and copyright system. Its patent office granted the DABUS patent in 2021, naming an artificial intelligence as inventor, the only such grant anywhere, and its copyright law has protected computer-generated works since 1992, assigning authorship to the person who arranged their creation. Both accommodations carry heavy caveats, the patent office examines only formalities and the computer-generated-works rule predates generative AI, but together they place South Africa nearer the open pole than any peer. The country also carries real weight abroad, holding the 2025 G20 presidency whose Johannesburg declaration launched an AI-for-Africa initiative, and it has a genuine domestic docket, with courts already sanctioning lawyers for AI-fabricated citations. For a jurisdiction whose AI-specific rules are policy, the IP posture and the international footprint are the profile's centre of gravity.

The site does three things: it maps every instrument that governs AI in South Africa, 16 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. South Africa's footprint is large, anchored by its 2025 G20 presidency and its BRICS membership.

Behind this country report

Muhammad Deckri AlgamarIP & digital law · LL.M. (IE-WIPO), AIGP

Contribute

Know of a ruling or proceeding in South Africa where AI is central? Send a source link; it will be reviewed before publishing.

Submit a case
An independent research and educational project. Information is provided as-is, may be incomplete, and does not constitute legal advice. Case entries record reported proceedings on a stated date, drawn from public sources, not a legal finding of guilt or liability.