AI Safety UAE

AI Safety UAE (aisafety.ae) is the UAE's first civil-society AI-safety community, co-founded by Wes Ezzeddine and Naiyarah Hussain as a hybrid enterprise-services and public-community initiative. Its aisafety.com description captures the mandate: to "catalyze collaboration, literacy, and governance frameworks that contribute to advanced AI systems being aligned with human and national interests." It is the corpus's first entry for any Middle East or North Africa (MENA) AI-safety local group, and the sole MENA entry in the AISafety.com global communities directory's 222-community span.

Founding and leadership

The organisation is co-founded by Wes Ezzeddine — a Dubai-based tech and AI-safety specialist who also facilitates with the European Network for AI Safety (ENAIS) — and Naiyarah Hussain, a senior engineer with prior roles at IBM, Lightning AI, and SurrealDB. Their international footprints anchor AI Safety UAE in the global EA-adjacent AI safety practitioner network rather than in purely local government-sector work. Two advisors supplement the founding team: Grace S. Thomson (AI policy strategy spanning MENA, Latin America, Europe, and the US) and Mohamed Yasser (software architect with 19 years of experience, specialising in government-sector clients in the region). Four working group leads divide the community's programme: Wes Ezzeddine (Literacy), Naiyarah Hussain (Convening & Coordination), Nikitha Ann Mathew (Governance & Regulation), and Omar Abdullah (Technology & Security).

Programme — enterprise services and community layer

AI Safety UAE runs a hybrid model, combining paid enterprise services with a free public community layer. On the enterprise side: foundational AI safety training sessions (2-hour live for non-technical and mixed audiences), intensive one- and four-week courses covering AI alignment, societal risk, ISO 42001 compliance, and technical safety for developers and GRC professionals, LLM and vision model audits with adversarial testing, and governance-framework development. Partners are BlueDot (the EA-adjacent AI safety course provider) and CodersHQ (the UAE government's national coder-community programme).

The community layer runs meetups, study groups, regional forums, stakeholder roundtables, and hackathons open to the public, with membership via the Luma calendar. Events have included a Trade & Commerce AI Governance Workshop planned for Dubai AI Week. The Luma community description encompasses "government entities, private companies, academics, and researchers" — a framing that reflects the UAE's top-down AI-governance landscape while signalling that the community is not confined to any single sector.

The four working groups operationalise the mandate: Literacy (study groups, certifications, expert guidance on AI policy), Governance & Regulation (policy research, stakeholder roundtables, framework development, and public-policy contributions), Technology & Security (AI safety tool-building, mentoring, technical research, and standards advancement), and Convening & Coordination (talent connection, communications, events, partnerships, and advocacy).

Place in the movement and civic context

AI Safety UAE is the corpus's first MENA entry — the only AI-safety community organisation across the Arab world in the AISafety.com directory. This reflects a real gap: a 2022 EA Forum mapping of technical AI safety in the UAE, produced by EA NYU Abu Dhabi, concluded that "we could not identify specific individuals or organizations in the country working on the problem" of technical AI safety. AI Safety UAE represents the community-building infrastructure that post-dates that scan.

The civic context warrants honest acknowledgement. The UAE constrains independent civil-society organising through a legal framework that limits freedom of association outside government-sanctioned channels. AI safety community work sits in a substantially more favourable position than political advocacy, however: the UAE's own agenda actively promotes AI leadership, governance capacity, and alignment with international safety frameworks — the country joined the Hiroshima AI Process, established a Ministry of State for Artificial Intelligence, and issued the UAE Charter for the Development and Use of AI in 2024. An AI safety community that builds governance literacy and practitioner capacity is aligned with, rather than in tension with, state priorities, affording it operational space that a rights-advocacy or surveillance-accountability group in the UAE would not have. The founders' international ENAIS connections and Wes Ezzeddine's speaker role at the international SUMMIT 333 AI Safety conference in Abu Dhabi (November 2025, alongside Max Tegmark and Jaan Tallinn) situate the group within the global civil-society AI safety movement rather than as a purely state-proximate entity.

Alongside Berlin AI Safety, AI Safety 東京, AI Safety South Africa, the Singapore AI Safety Hub (SASH), and AI Safety India, AI Safety UAE extends the corpus's coverage of EA-adjacent AI-safety community-building groups across a new geo. What distinguishes AI Safety UAE from those counterparts is its hybrid commercial/community structure: where Berlin and Tokyo began as volunteer-run meetup formats and Singapore launched as a nonprofit co-working space, AI Safety UAE pairs community-building with revenue-generating enterprise services — training, audits, and governance frameworks — a model that reflects the UAE's high-cost expatriate professional economy, where self-sustaining community organisations often require a commercial layer to operate sustainably.