Responsible AI & Trust
Human oversight, transparency and evidence are built into how we make AI — so our partners can check our work, not take it on faith.
Code Hub Artificial Intelligence Developing Services L.L.C · Commercial Licence 768066 · Licensed in Dubai since October 2016
Illustration: six arcs of light rise around a human check mark at the centre, one arc for each commitment in our responsible AI charter — human oversight, transparency, fairness, privacy, accountability and sustainability.
Responsible AI
Six commitments describe how we work. We map each to the Dubai AI Ethics Principles and, where one applies, the UAE AI Charter.
AI drafts. People decide. Experts stay the authority.HumanityEthicsHuman oversightGovernance and accountability Maps to Dubai AI Ethics Principles: Humanity, Ethics; UAE AI Charter: Human oversight, Governance and accountability.
We say how AI is used — and label what is illustrative.EthicsTransparency Maps to Dubai AI Ethics Principles: Ethics; UAE AI Charter: Transparency.
Learners progress on demonstrated evidence, never on time served.EthicsInclusivenessAlgorithmic bias Maps to Dubai AI Ethics Principles: Ethics, Inclusiveness; UAE AI Charter: Algorithmic bias.
Learner data trains no external foundation models.SecurityData privacySafety Maps to Dubai AI Ethics Principles: Security; UAE AI Charter: Data privacy, Safety.
Decisions leave a record that can be traced and reviewed.EthicsSecurityGovernance and accountability Maps to Dubai AI Ethics Principles: Ethics, Security; UAE AI Charter: Governance and accountability.
Right-sized models, efficient architectures, systems built to last.HumanityUAE Net Zero by 2050 Strategic Initiative Maps to Dubai AI Ethics Principles: Humanity; National priority: UAE Net Zero by 2050 Strategic Initiative.
Dubai AI Ethics PrinciplesDubai AI Ethics Principles & Guidelines
Its four principles: Ethics · Security · Humanity · Inclusiveness
UAE AI CharterUAE Charter for the Development and Use of Artificial Intelligence (2024)
Principles referenced here: Human oversight · Safety · Algorithmic bias · Data privacy · Transparency · Governance and accountability
The mapping is ours. It is not an assessment or endorsement by either framework's issuer.
Human oversight, in practice
In Maestro (opens in a new tab) Studio, AI engineers the course from your learning outcomes — and your subject-matter experts approve every academic decision before it ships.
Nothing reaches a learner unreviewed. Illustration: AI-drafted lessons, assessments and feedback travel to a human reviewer, who approves each one on to learners or returns it for revision.
No academic decision is made by AI alone.
Illustrative interface1
The human gate
No academic decision is made by AI alone.
Documented model use
How AI models are used is documented.
Auditability of adaptation
Prompt versions are immutable, and there is an audit log.
Learner data boundaries
Learner data trains no external foundation models.
Product security
Only what each product's own documentation states, with a link so you can check it.
Maestro
The AI-native adaptive learning system
THE-CODE API
Multiple AI providers. One API.
The UAE's data protection law
The UAE's federal law on the protection of personal data, in force since 2 January 2022. Our privacy notice explains how this site handles your enquiries under it.
Read our privacy noticeResponsible compute
The greenest compute is the compute you never spend. We size models to the task, and build systems that last.
Illustration in three steps. First, one oversized model lights a whole grid of compute for a small task. Then the grid contracts to a small, right-sized cluster that does the same job. Finally, a long timeline shows that cluster improved in place over the years instead of being rebuilt.
The largest model for every task, however small.
Same job. The model sized to the task.
Improved in place — not rebuilt.
Illustrative. Not to scale.
One size for everything
The largest model for every task, however small.
Right-sized
Same job. The model sized to the task.
Built to last
Improved in place — not rebuilt.
Illustrative. Not to scale.
The smallest model that does the job well — not the largest by default.
Designs that do more with less, from the first line of code.
Compute spent where it adds value, and nowhere else.
Built to last and improved in place, rather than rebuilt from scratch.
Designed with the UAE's sustainability priorities in mind
Registered in Dubai
The facts a reviewer needs to verify us — our registered details, from our Dubai commercial licence.
Build with us
Bring us your hardest questions about oversight, data and security. We'll answer plainly, with sources you can check.