Comparing ChatGPT and Claude: Legal and Policy Considerations for Samoa

Understanding Samoa’s Evolving Legal and Policy Landscape for AI

Artificial intelligence (AI) is gaining traction worldwide, and Samoa is no exception. While there is growing interest in generative AI tools such as ChatGPT and Claude, Samoa does not yet have a dedicated AI statute or regulatory framework. Instead, AI use is governed indirectly through existing digital government policies, information and communication technology (ICT) regulation, public law duties, and related statutes such as defamation law.

The Ministry of Communications and Information Technology (MCIT) and the Office of the Regulator are leading efforts in digital transformation. Initiatives like Digital Samoa and the ICT Sector Plan 2022–2027 emphasise privacy, security, and legal reform as policy priorities. Recent UNDP workshops have further highlighted the importance of human rights, equity, and data privacy in Samoa’s approach to AI.

Given this context, drafting an article comparing ChatGPT and Claude should be framed against an evolving regulatory backdrop, where the legal landscape is developing but not yet settled.

Comparing ChatGPT and Claude: Features with Legal Significance

A simple feature-by-feature comparison misses the deeper governance relevance of generative AI systems. For Samoa, the most critical distinctions between ChatGPT (OpenAI) and Claude (Anthropic) relate to:

- Data handling and privacy controls

- Auditability and administrative oversight

- Multilingual capabilities

- Enterprise controls and deployment options

- Regional data hosting

OpenAI’s API data usage policies and enterprise solutions offer SAML SSO, role-based access, analytics, and in some cases, regional data residency (e.g., Singapore, Australia). These features are highly relevant for Samoan institutions wishing to manage compliance risk or ensure suitable data governance.

Anthropic’s Claude documentation highlights audit logging, export controls, and multilingual support. In a Pacific context, strong multilingual capability is particularly important for preserving language inclusion in public services and education. Claude also offers enterprise tools for oversight, though fewer details are publicly documented compared to OpenAI.

For government agencies, schools, and banks in Samoa, these differences are not merely technical—they can affect procurement decisions, legal compliance, and defensibility in case of AI-related incidents.

Data Governance, Privacy, and Cross-Border Risks

One of the most pressing issues for Samoa is data governance. As the MCIT Annual Report 2022–2023 and the ICT Sector Plan highlight, Samoa is still developing its privacy and data protection framework. This raises the practical importance of the technical and contractual safeguards provided by AI vendors.

Key risks include:

- Personal or confidential data entered into AI prompts may be retained, logged, or transferred overseas

- Use of overseas cloud hosting can create uncertainty about foreign access to Samoan data

- Public sector use requires robust auditability, retention controls, and incident response capabilities

OpenAI provides some regional data hosting options, which may help address data sovereignty concerns for Samoan users. Anthropic’s documentation on audit logging is relevant for maintaining institutional accountability. In the absence of comprehensive domestic privacy law, these technical and contractual controls serve as the main safeguards.

Liability, Misinformation, and Public Sector Risk

Generative AI brings familiar legal risks in new forms. The Defamation Act 1993 (Samoa) is particularly relevant, as it applies to both individuals and government entities. If an AI system produces a false statement that is published by a public servant, liability may attach to the agency.

Other risks include:

1. Defamation: False allegations or fabricated content generated by AI
2. Negligence: Failure to verify AI outputs before public release
3. Confidentiality breaches: Sensitive data entered into AI prompts
4. Procedural unfairness: AI-influenced decisions without adequate human oversight

Statutory immunity generally depends on acts being done "honestly and without negligence" (see Land Transport Authority Act 2007). As such, agencies must ensure robust human supervision, transparent recordkeeping, and reasonable oversight when deploying AI systems.

Whether using ChatGPT or Claude, the decisive legal factor is often not the model itself but the adequacy of deployment controls—such as audit logs, admin restrictions, and retention management.

Policy Implications for Government, Education, Language, and Inclusion

Samoa’s broader development agenda, as outlined in Samoa 2040, places a premium on digital literacy, inclusive growth, and equitable access. While there is no dedicated generative AI policy for education or public administration, several practical implications arise:

Education

• Academic integrity and teacher oversight are essential.
• AI systems should support equitable access, especially in multilingual or lower-resource contexts.
• If one model (such as Claude) better supports Samoan or Pacific languages, this could enhance fairness for students.

Government

• Agencies need clear procurement standards, human-review protocols, and restrictions on AI as a final decision-maker.
• Deployment should prioritise auditability, traceability, and records management.

Language and Culture

• Multilingual performance is not just a technical feature, but a governance issue.
• AI tools that better support Samoan can help bridge digital divides.

Digital Inclusion

• Systems requiring high-end infrastructure or costly enterprise agreements may exclude smaller organisations.
• Policy should focus on affordability and capability-building.

Practical Guidance for Samoan Lawyers, Professionals, and Agencies

Given Samoa’s evolving legal environment, the following practical steps are recommended when considering generative AI adoption:

- Assess vendor documentation: Scrutinise privacy, data retention, and auditability features offered by ChatGPT and Claude.

- Prioritise enterprise controls: Choose systems that allow for strong administrative oversight, regional data hosting, and audit logging.

- Develop internal guidance: Establish rules for human review, records management, and prompt confidentiality within your organisation.

- Monitor legal reforms: Stay up to date with changes to digital policy, privacy law, and AI governance in Samoa.

- Balance innovation with inclusion: Ensure AI adoption supports digital equity, especially for Samoan language users and less-connected communities.

Looking Ahead: Building Trustworthy AI Use in Samoa

Samoa’s approach to AI is defined by ongoing legal and policy reform, a commitment to digital inclusion, and the practical realities of a small island state. In this context, the distinctions between ChatGPT and Claude are more than technical—they shape how organisations manage privacy, accountability, and risk.

As Samoa continues to develop its digital governance framework, public sector bodies, educational institutions, and legal professionals should focus on responsible procurement, robust oversight, and context-appropriate safeguards. By grounding AI adoption in clear internal policy and informed vendor choice, Samoa can advance digital transformation while protecting public trust and equity.