What is AI regulation in Tonga?
AI regulation: countries and regions
Tonga has no dedicated AI law or regulatory framework. Artificial intelligence is governed under existing technology and data laws, such as the new Privacy Act and communications regulations, rather than by a specific statute. The country's digital strategy (Digital Government Framework 2019-24) and participation in regional initiatives (like the Pacific Digital Democracy project) provide context, and Tonga follows international AI ethics principles (UNESCO's recommendation) as guidelines. No Tongan regulator or law is focused solely on AI.
Reviewed by Jackie, Head of Learning & Development, Levellers · Last reviewed 8 June 2026
What this means
In plain terms, Tonga has not passed any law or policy specifically about artificial intelligence. Instead, AI systems in Tonga fall under the same laws that apply to other technologies. For example, if an AI service handles personal data, it must comply with Tonga's data protection rules. If it involves financial transactions or telecommunications, the relevant sector laws apply. Essentially, Tonga treats AI as part of its broader digital landscape without separate AI legislation.
Tonga's government has worked on digital transformation but has not singled out AI. Its Digital Government Framework (2019-2024) focuses on e-government and ICT services, not AI per se. Tonga has updated other laws - for example, it passed a comprehensive Privacy Act in 2025 and a new Cybercrime law - but these cover data and cyber offences generally, not AI specifically. The Privacy Act will govern how personal information is used (affecting any AI that processes data), but it doesn't impose new rules just for AI.
Tonga also adheres to international norms on AI. As a member of UNESCO and the UN, it has endorsed the UNESCO Recommendation on AI Ethics, which stresses human rights, transparency and oversight. Regionally, Tonga participates in Pacific initiatives on digital governance (for example, the EU/UNDP Pacific Digital Democracy Project). These guidelines and projects shape policy thinking, but they do not create laws in Tonga. In summary, there is no separate AI law or regulator in Tonga; AI is managed through the general legal and policy framework for technology and data.
Why it matters
The practical effect is that organisations using AI in Tonga must rely on general laws and self-regulation. There are no specific Tongan rules on AI safety, bias or transparency. This matters because developers and users of AI still need to manage risks on their own. For example, a business deploying AI must ensure it does not breach privacy or mislead consumers even though no AI law explicitly requires it. Similarly, government agencies using AI should apply ethical principles (such as fairness and accountability from international guidelines) in the absence of local mandates. On the upside, the lack of rigid AI regulation means less compliance burden right now, but it also means unpredictability. Companies and public leaders should keep up with global AI standards and be ready for future laws, because doing nothing is not a safe option. Proactively following best practices (like those in the UNESCO and OECD AI guidelines) and preparing for the Privacy Act's enforcement are prudent steps when specific AI rules are not yet in place.
How it works
AI-specific Legislation
Currently, Tonga has enacted no laws that mention AI explicitly. There is no "AI Act" or standalone legal framework for artificial intelligence. All AI applications are therefore treated under the broad scope of existing laws. For example, if an AI system is used to make business decisions or recommendations, Tonga has no special AI requirements for transparency or risk assessment - only the usual commercial, civil or criminal laws would apply if something goes wrong.
Data Protection and Privacy
In 2025 Tonga passed its first comprehensive Privacy Act, which will regulate personal data processing. This new law sets up a Privacy Commission and rules for how any organization (including those using AI) must handle personal information. Until it comes into force (by proclamation) there is no statutory data protection in Tonga. DLA Piper notes that previously Tonga relied on common law for privacy. When active, the Privacy Act will apply to AI systems that collect or use personal data, but it does not create any AI-specific obligations beyond general data protection standards.
Sectoral Regulation
Tonga's approach to new technologies is largely sectoral. The central bank, for example, runs a FinTech Regulatory Sandbox (launched in June 2025) to let companies test innovative financial services. This sandbox can include AI-powered fintech products, but it is a broad innovation framework, not an AI-only rule. Other sectors follow their usual rules: the Telecommunications Act and Tonga Communications Commission govern telecom networks (without special AI provisions), and the Companies and Foreign Investment Acts cover business registration. In practice, AI systems in Tonga must comply with the rules of the industry they serve, with no additional AI-specific regime.
Regional and International Alignment
Regionally, Tonga participates in Pacific digital governance initiatives (such as the Pacific Digital Democracy Project) that promote responsible use of technology. There is no Pacific-wide AI law; instead Pacific governments share best practices and harmonize where possible. Internationally, Tonga is bound by its UN and UNESCO commitments. The 2021 UNESCO Recommendation on AI Ethics - applicable to all member states - emphasises protecting rights, transparency and human oversight in AI. Tonga has endorsed those principles, but they remain guidelines rather than enforceable laws. No international treaty or UN regulation specifically covers AI in Tonga, although global norms can influence Tonga's future policy decisions.
Examples
- **FinTech AI innovation:** A Tongan startup with an AI-powered payments app can apply to the National Reserve Bank's FinTech Regulatory Sandbox (launched June 2025) to trial its service under supervision. The sandbox imposes clear testing limits and safeguards. Even in this trial phase, the company must still follow Tonga's existing laws (for instance, consumer protection and any future privacy rules), but it benefits from a structured environment to validate its AI technology before full deployment. - **Government AI deployment:** If a Tongan government agency builds or uses an AI system (for example, to analyse health or climate data), it must operate within current laws. The agency would be required to protect any personal data according to the Privacy Act 2025 once it comes into force. In practice, the agency should also follow transparency and fairness principles (as advised by UNESCO's AI ethics recommendations) even though no local law explicitly mandates that. This ensures public trust until specific regulations might be introduced.
Common misunderstandings
- **"Tonga's Privacy Act is an AI law."** In fact, the Privacy Act 2025 regulates personal data for all technologies. It does not contain any provisions specific to AI algorithms or decision-making. It simply means any AI system handling personal data must obey the same rules as other data processors. - **"UNESCO's AI ethics rules are binding here."** Some assume that because Tonga agreed to UNESCO's AI ethics recommendation, it now has AI rules. In reality, that recommendation is voluntary guidance for governments. Tonga has endorsed its principles (like human rights and accountability), but following them is not legally required. - **"Without an AI law, AI is illegal in Tonga."** The absence of a special AI statute does not ban AI. Organisations can develop and use AI as they wish; they just must respect existing laws (for example, they cannot defraud customers or misuse data). There is no Tongan regulation that criminalises AI usage itself. - **"The Digital Government Framework regulates AI."** Tonga's 2019-2024 Digital Government Framework drives ICT modernization and e-government. It does not establish any technology standards or rules for AI. It sets general goals (like improving services online) but imposes no specific AI requirements. - **"FinTech sandbox restricts AI."** The Reserve Bank's sandbox is sometimes seen as a control on innovation. In fact, it is an enabling mechanism: it allows fintech firms (possibly using AI) to experiment under oversight. Participation is optional and meant to guide safe innovation, not to ban technology.
Risks and boundaries
- **No special enforcement:** Currently there is no Tongan agency or law that enforces AI-specific rules. The risk is assuming one exists. Any wrongdoing by an AI system (for example, an AI that inadvertently spreads false information) would be handled under existing laws (such as defamation or fraud statutes). There are no extra AI penalties or licences in Tonga, so users and developers must rely on general legal boundaries. - **Data law delay:** Although the Privacy Act 2025 will govern personal data use, it has not yet commenced by proclamation. This means there is a gap: personal data (including that used by AI) currently lacks statutory protection. Organisations should be cautious in the interim. Once the law is active, non-compliance could lead to enforcement actions. - **International limits:** Tonga's approach is aligned with global principles but not controlled by them. UNESCO's guidelines and UN frameworks provide ethical guidance, but they do not override Tongan law or create penalties here. Similarly, foreign AI laws (for example, the EU AI Act) do not apply in Tonga unless Tongan entities do business in those jurisdictions. Tonga's own laws apply only to activities within its territory or involving Tongan data subjects. - **Sector-specific scope:** Any current "limits" on AI in Tonga come from sector laws. For instance, the FinTech sandbox and related financial regulations only cover firms that join that program. Outside of those programs, AI developers have flexibility. This broad scope is a boundary: Tonga's rules don't extend beyond the existing legal framework, so novel AI scenarios might slip through regulatory gaps. - **Misapplied expectations:** Because there is no AI law, there might be confusion with unrelated laws. For example, media or content regulations do not automatically govern AI-generated content. Businesses should not assume that AI is independently regulated - instead, they should focus on compliance with familiar laws (privacy, consumer protection, etc.) as the true legal boundary.
What to do next
- **Review data governance now:** Organisations developing or deploying AI should immediately assess their data handling against Tonga's laws. They should prepare for the Privacy Act 2025 by mapping data flows, obtaining proper consent and strengthening security. Even without an AI law, strong data governance will mitigate future risk. - **Adopt ethical best practices:** Leaders can follow international AI ethics and risk management guidelines voluntarily. For example, they should ensure any AI system is transparent, fair and explainable, as recommended by UNESCO. Creating an internal AI ethics code or conducting impact reviews can demonstrate responsibility in the absence of regulation. - **Engage with regulators and programs:** Fintech firms should consider participating in the Reserve Bank's sandbox to safely test AI innovations in finance. Similarly, companies in other sectors should consult their relevant regulators (e.g. telecommunications, health) to clarify compliance. Keeping an open dialogue with regulators helps prepare for any future rules. - **Monitor policy developments:** Policymakers and business leaders should keep an eye on global AI trends and potential new laws. They might form an inter-agency AI working group or develop a national AI strategy as a next step. In the meantime, aligning with regional initiatives (like Pacific digital projects) can help ensure Tonga's interests are represented in any future AI governance frameworks. - **Educate and raise awareness:** Finally, governments and organizations should train staff on AI concepts and risks. Building AI literacy and ethical awareness will put Tonga ahead when formal AI policies do arrive. Proactive learning and stakeholder consultations are practical ways to stay ready in this evolving field.
Have a question or a suggestion, or want to understand how we research and review these guides? Read about our editorial standards and how to reach us.
FAQs
Is there a specific law or regulator for AI in Tonga?
No. Tonga has not enacted any law that specifically governs AI. There is no dedicated AI regulator. Instead, AI is covered under general laws (like data protection and sector laws).
How is personal data used by AI protected in Tonga?
Tonga passed a Privacy Act 2025 to regulate personal data processing, but it is not yet in force. Once active, this law will apply to any technology (including AI) that handles personal information. Until then, Tonga relies on common law principles for data protection.
Do UNESCO or UN AI guidelines affect Tonga's laws?
Tonga endorses UNESCO's Recommendation on AI Ethics, but that guidance is not legally binding. The Recommendation provides ethical principles (e.g. human rights, transparency) that Tonga can follow, but it does not override Tongan law.
What if an AI system breaks a law in Tonga?
Any harm caused by AI (such as misinformation or bias) would be dealt with under Tonga's existing laws (e.g. criminal or civil statutes). There are no extra "AI-specific" penalties, so liability would depend on which general law was violated.
Can companies test AI innovations in Tonga?
Yes. For fintech applications, the National Reserve Bank offers a regulatory sandbox (launched June 2025) where companies can trial new services including AI under oversight. Outside the sandbox, companies may deploy AI as long as they comply with general regulations.
Will Tonga create an AI law soon?
There is no announced AI bill at this time. Leaders should watch for updates, but for now Tonga's approach is to use existing legal frameworks. Organizations should focus on compliance with those frameworks and best practices, and prepare for any future AI policy developments.