The Complete Guide to Using AI in the Government Industry in Malta in 2025

Why AI matters for government in Malta in 2025 is straightforward: the Malta Digital Innovation Authority is realigning the national Malta AI Strategy and Vision 2030 while EU-wide rules such as the AI Act and GDPR set strict boundaries, so public teams face a once-in-a-generation chance to boost services safely - from traffic‑management pilots designed

to reduce congestion and emissions

to CPSU forecasting for medicines and EU‑funded radiology projects like

BreastScan

that promise faster screenings.

The legal backbone and sector guidance are mapped out in the detailed AI practice guide for Malta, which flags procurement, liability and data risks that every procurement officer and policy team must heed.

Practical upskilling is essential: hands-on courses such as Nucamp's AI Essentials for Work bootcamp teach prompts, tools and workplace use-cases so civil servants can implement AI projects that cut costs without cutting corners - imagine traffic lights that learn to clear rush‑hour bottlenecks and immediately cut idling and emissions.

Malta's national plan - published as the “Strategy and Vision for Artificial Intelligence in Malta 2030” - purposefully positions the islands as an innovation launchpad, built around three strategic pillars: investment and start‑ups, public‑sector adoption and private‑sector uptake, supported by cross‑cutting enablers in education and workforce development, ethics and law, and ecosystem infrastructure.

The Malta Digital Innovation Authority (MDIA) leads a realignment of the Strategy through 2025, using intensive stakeholder consultations and public feedback to keep the roadmap current; the OECD summary highlights that the Strategy also backs six pilots across health, education, traffic, customer service, tourism and utilities to demonstrate citizen-facing gains.

Practical features include a national AI certification programme and actions to support Maltese‑language models, data availability and access to compute - and the Strategy even treats Malta's compact geography as an advantage, small enough to run national pilots that can be scaled abroad.

For the official text see the MDIA's Malta AI Strategy and Vision 2030 and the OECD's overview of “The Ultimate AI Launchpad” for further detail.

“…leverage its natural resources and size, as well as innovative public policy, to translate a bold leadership vision into a set of tools, incentives, resources and collaborative ecosystems that accelerate the journey from AI development to AI adoption, leading to commercial success, social benefit and international recognition.”

Malta's regulatory architecture for AI in 2025 is a layered, cooperative system built around the Malta Digital Innovation Authority (MDIA) as the lead implementer of the EU AI Act, working hand‑in‑glove with sectoral supervisors such as the MFSA and MGA to marry EU‑level rules with industry realities; the MDIA's sandboxes and voluntary certification pathways let innovators test systems against DORA, NIS2 and the AI Act before full deployment, while sector regulators translate those obligations into industry‑specific guidance for finance, gaming, transport and health.

At the same time the Office of the Information and Data Protection Commissioner (IDPC) has been elevated to a central watchdog role - designated to protect fundamental rights and act as a Market Surveillance Authority - so that privacy, automated‑decision rules and high‑risk uses (think border biometrics or justice‑related systems) receive targeted oversight.

The result is a practical, risk‑based ecosystem: MDIA-led assurance and sandboxes to foster safe innovation, plus specialized supervisors ready to audit suppliers and require back‑to‑back contractual safeguards, creating a single national loop from development to market surveillance that aims to keep Maltese pilots scalable and compliant as they move from Valletta to wider European markets (MDIA leads Malta's implementation of the EU AI Act, Chambers' Malta AI practice guide).

“we are already ‘preparing for this responsibility by ensuring that we introduce the necessary legislative amendments and build the required expertize.'”

In 2025 Malta's AI activity is concrete and citizen-facing: a government press release highlights that Malta has strong public digital service performance, while pilots and use-cases focus on tourism, justice and service efficiency - think a tourist digital identity that offers sustainable recommendations to nudge climate-friendly travel and personalised tourism features that reduce call-centre load (Malta tourist digital identity and sustainable travel recommendations pilot), alongside paralegal document automation reshaping public legal offices and creating legal-tech workflows (paralegal document automation and legal-tech workflows in Malta).

These projects sit alongside Malta's Open Government commitments (2023–2025), which prioritise a justice forum, better access to services for victims, clearer information about rights, and youth-proofing legislation - though civil society has flagged that the co-creation process did not always allow sufficient time to influence the commitments, a reminder that inclusive design matters as pilots scale from Valletta to wider national rollout.

Malta “outperformed the European average by 18% in the area of public digital services” - Office of the Principal Permanent Secretary press release on Malta's public digital services performance

Key bodies and people to watch in Malta's 2025 AI scene cluster around a compact, well‑connected hub: the Malta Digital Innovation Authority (MDIA) is the national lead for implementing the EU AI Act and supporting trustworthy, human‑centric AI, with a small team based out of Birkirkara's Twenty20 Business Centre and visible programmes for upskilling and startups (MDIA Artificial Intelligence services); the MDIA also runs the DiHubMT European Digital Innovation Hub, which offers

test before invest

services, high‑performance computing and AI testbeds to SMEs and public bodies and is coordinated by Bernard Montebello (di hubmt@mdia.gov.mt) - see the Malta‑EDIH entry for service details (Malta EDIH (DiHubMT) service details).

Practitioners and procurement teams should note the MDIA's Technology Assurance Sandbox - a practical residency where applicants can choose which controls apply at each stage, gain technological assurance and smooth the path to MDIA recognition - a real advantage for small teams wanting to pilot nationally scalable systems without being swamped by compliance work.

For quick orientation, remember three touchpoints: regulatory guidance from the MDIA, the DiHubMT testbed for pilots, and the sandbox for early validation; together they form the pragmatic back‑bone for Malta's AI rollout, tangible enough that even a start‑up can demo in a certified environment before seeking buyers.

Generative AI tools are rapidly shifting from back‑office helpers to decision‑support engines, but in Malta they must sit squarely inside the GDPR‑based framework transposed by the Data Protection Act 2018 - meaning legal bases, transparency, data minimisation and robust security are non‑negotiable (Maltese Data Protection overview under the Data Protection Act 2018).

Models trained on personal or sensitive data (think health, biometric or CCTV feeds) will often trigger a DPIA and, where public‑interest processing is involved, prior consultation with the Office of the Information and Data Protection Commissioner; breach notification timelines and DPO rules also apply.

At EU level the EDPB's recent work on “AI models” underlines how GDPR principles (rights to information, human review of automated decisions, and strong technical measures like pseudonymisation) must be woven into model lifecycle governance (EDPB guidance on AI models and GDPR compliance).

Contracting and IP issues matter too: controllers remain responsible for compliance and should insist on clear processor clauses, retention limits and vendor representations about training data and model behaviour to manage both data‑protection and ownership risks - practical counsel on these intersections is available in national GDPR implementation guides (Malta national GDPR implementation guide).

Picture a traffic pilot in Valletta: without pseudonymisation, a DPIA and tight vendor contracts, an otherwise promising generative model can become a compliance headache faster than a rush‑hour bottleneck clears.

Liability, standards and procurement for Malta's government AI projects sit at the intersection of old-school fault rules and new EU obligations: domestic law still relies on the Civil Code's fault-based tort framework (articles 1031–1033) so deployers and developers - not the machine - are the likeliest defendants if an AI system causes harm, yet the EU has already tightened the net with a revised Product Liability Directive that extends strict liability to software even as Malta works through transposition and the AI Liability Directive was dropped; practical procurement therefore must bake in back‑to‑back contractual safeguards, clear processor clauses, warranties about training data and model behaviour, and DORA/NIS2-aware outsourcing terms so that public buyers can call on suppliers if a predictive model misfires.

The MDIA's sandboxes and voluntary certification pathways help teams test conformity before wide release, and sectoral guidance (for example from the MFSA) steers financial and critical‑service procurements; think of a driverless‑bus pilot - the legal checklist (DPIA, fail‑safes, vendor liabilities, insurance) must be resolved before the bus leaves the depot.

For practical legal background see the detailed GLI country chapter and Chambers' Malta AI practice guide for procurement and liability pointers.

Source: AI, Machine Learning & Big Data Laws 2025 - Malta (Global Legal Insights) and Artificial Intelligence 2025 - Malta (Chambers)

Malta's 2025 push to build AI capacity pairs practical upskilling with targeted cash - led by the MDIA's Applied Research Grant (MARG), which explicitly backs AI, digital trust, data representation, sustainability and other emerging tech and has already supported nine projects with awards up to €40,000; the scheme is designed to move ideas from university labs into citizen‑facing pilots and commercial uptake, and applications are open annually via the MDIA portal (MDIA Applied Research Grant (MARG) portal).

Complementary national instruments and rebranded R&I bodies (Xjenza/MCST) and university support channels broaden the pipeline, while an April 2024 government announcement highlighted a separate €1.2 million two‑year funding programme to seed digital projects - signalling that Malta is layering small, focused grants to create an ecosystem where a Smart Valletta pilot or a spoken‑Maltese corpus can get traction without waiting for big venture rounds (OECD overview of the MDIA Applied Research Grant (MARG) and the national funding call).

“Today's initiatives reflect the intensive efforts underway to ensure that Malta remains well-prepared for the future, as technology continues to evolve at an accelerated pace.” - MDIA CEO Kenneth Brincat

Start small, follow the rules, and build for scale: Maltese government teams should first map each use‑case against the EU AI Act risk categories and the MDIA's guidance so it's clear whether a conformity assessment, DPIA or prior checks are required, then move promising pilots into the MDIA's Technology Assurance Sandbox or DiHubMT testbeds to “test before invest” and iron out controls without slowing delivery (MDIA Artificial Intelligence services and guidance); procurement teams must embed back‑to‑back contractual safeguards, processor clauses and vendor warranties (training‑data provenance, retention limits, explainability SLAs) so deployers can call suppliers to account, and where critical services are involved ensure DORA/NIS2 resilience checks are included.

Pair every technical rollout with GDPR‑aligned measures (DPIA, anonymisation/pseudonymisation, human oversight and clear user notices), document governance in a living impact assessment, and use small grants or applied‑research funding to fund pilot validation before scaling - this mirrors Malta's strategy to prioritise public‑sector adoption and citizen‑facing pilots in the national AI roadmap (Malta AI Strategy 2030 (OECD summary)).

The practical payoff is tangible: a Valletta traffic pilot tested in the sandbox can tune lights to cut idling and emissions while keeping CCTV data pseudonymised and vendor liabilities crystal clear, turning legal risk into measurable public benefit.

“Malta aspires to become a hub of artificial intelligence implementation. The ambition is… to provide the best possible environment for this technology to flourish,” he said.

Next steps for Maltese government teams are practical and tightly focused: map each use‑case against the MDIA‑aligned risk categories and move high‑value pilots into the MDIA's sandboxes and certification pathways so compliance is baked in from day one (see the Malta AI Strategy and Vision 2030), then seek small, targeted funding to validate outcomes - the MDIA Applied Research Grant (MARG) is designed to seed citizen‑facing pilots and proof‑of‑concepts before scale (MDIA Applied Research Grant (MARG)).

Pair legal and procurement safeguards (DPIAs, processor clauses, vendor warranties and fail‑safe requirements) with hands‑on staff training so teams can turn policy into production without surprises: local, instructor‑led courses cover policy automation, fraud detection and service redesign while short practical bootcamps teach prompt craft and workplace integration - for example, Nucamp's AI Essentials for Work bootcamp trains non‑technical staff to use AI tools safely.

Use Malta's compact geography as an advantage - run a national pilot end‑to‑end, learn fast, tweak controls, and only then scale - and keep stakeholder co‑design central so technological wins translate into trusted public benefit rather than friction for citizens.