Built different,
on purpose.

Preamble: Our Founding Purpose

This constitution is made for the purpose to guide our work in developing and deploying our own unique artificial intelligence solutions that serve humanity, mitigate environmental impact of AI, and uphold the highest standards of equity and social justice.

As an Australian company, we recognise our responsibility to operate within the framework of Australian law, to respect the sovereignty and teachings of First Nations peoples, whilst contributing to the wellbeing of our local communities and the broader global society.

We believe that artificial intelligence should be a force for good, further improving human capability, and ensuring that the benefits of technological progress are shared equally across throughout society.

Ethical AI Principles

Transparency: All AI systems and solutions from Pockel must be designed to be explainable and interpretable. Committing to clearly communicating how our AI makes decisions, the data it uses, and its limitations. Black-box algorithms that cannot be audited have no place in our work.

Fairness and Non-Discrimination: We must always proactively look for and mitigate bias in our systems. They must not discriminate on any basis. We will employ diverse datasets and regular fairness audits to ensure equitable outcomes.

Human Oversight: AI is a tool to empower human decision-making, not replace it. We require genuine human oversight for all of our steps, ensuring that critical decisions remain in human hands.

Accountability: We accept our role in ensuring AI is safe, ethical and meets the needs of our users, and take it upon ourselves to ensure it will remain so.

Environmental Responsibility

Sustainable Computing: We commit to minimising the environmental impact of our systems and operations. This includes selecting energy-efficient model architectures and potentially offsetting unavoidable emissions through verified carbon credits where possible.

Resource Efficiency: We will always prioritise smaller, more efficient models where they can deliver equivalent performance. We reject the common western "bigger is always better" paradigm and instead pursue the right-sized model for each application, reducing computational waste and environmental impact.

Equitable Access and Inclusion

Digital Inclusion: We work to bridge the digital divide by designing for low bandwidth environments and ensuring our pricing models and underlying never leave economically disadvantaged peoples behind.

Benefit Sharing: The economic value created by our systems must be shared with the communities and workers contributing to and affected by our systems. Rejecting extractive models concentrating wealth while displacing labour without support.

Open-Source Commitment: Where possible and appropriate, we will release tools, datasets, and research that can advance the public good within the balance of our commercial viability.

Australian Company

Cultural Safety: Our AI systems are designed to be culturally safe and respectful. We must always ensure our technology does not perpetuate cultural harm.

Local Employment and Supply Chain: We prioritise Australian-based talent and suppliers, contributing to our local economy.

Data Privacy and Security

Privacy by Design: Privacy is embedded into our products. We empower users the control over their own data, it is their right.

Security Excellence: Security is not a feature; it is a foundation.

Data Minimisation: We challenge every request for personal data with the question: is this strictly necessary? We do not hoard data for speculative future use.

User Control: Individuals have the right to control their own data, it is an extension of their person and hence they maintain their autonomy over it.

Governance and Accountability

Ethics: Pockel AI shall maintain Ethics counselling with diverse representation, including members with expertise in a variety of fields, with the authority to review projects that violate this may Constitution.

Stakeholder Engagement: We must maintain ongoing dialogue with affected communities, workers, customers, and organisations. Major product decisions require consultation, and feedback.

Annual Constitutional Review: Our constitution is a living document. Amendments may be needed to ensure it remains relevant and effective in a rapidly evolving technological landscape, and we must make them in good faith to our community.

Innovation with Integrity

Harm Avoidance: We will never develop AI systems designed to cause harm ever. Our innovation is bounded by our values. This is never for debate.

Dual-Use Diligence: For technologies with potential dual-use applications, risk assessments and safeguards are activated. When we cannot mitigate a risk, we do not proceed.

Truthfulness: We commit to honest representation of our AI capabilities. Never engaging in hype or misrepresentation that could lead to misplaced trust or over-reliance on incapable systems.

Collaborative Progress: We believe that responsible AI development is a collective endeavour. We must share safety research, whilst contributing to the development of beneficial AI norms and practices.