Who can claim the UK's free AI and technology training—and what does it prove?
The Government Digital Service has opened more than 200 free learning and certification pathways to eligible UK civil and public servants until 31 December. Last year's programme recorded over 10,000 registrations and 1,200 certifications, but the government has not published completion, skill-gain or workplace-outcome data.
The short version
- What happened
- Eligible civil and public servants can use more than 200 learning and certification pathways from 1 October to 31 December; contractors are excluded.
- Who it affects
- Eligible public servants can access training that might otherwise require departmental or personal funding, including beginner routes.
- What remains unproven
- Both sources are programme publications from the UK government; there is no independent evaluation of learning or workplace impact.
Keep in mind
Both sources are programme publications from the UK government; there is no independent evaluation of learning or workplace impact.
The 2025 registration and certification totals do not disclose the learner-level denominator needed for a completion or pass rate.
The sources do not separate AI-course participation from the other seven technology categories.
Provider-specific availability, prerequisites, voucher limits and redemption deadlines may differ from the umbrella campaign window.
Source analysis · 2 linked sources. Primary analysed source date: 1 October 2026.
Read the full report · 6 min The evidence, methods, implications and limitations
The offer is broad, but eligibility has a hard boundary
The UK Government Digital Service opened Get Tech Certified 2026 on 1 October for people working across the civil service and wider public sector. The hub lists more than 200 free learning and certification pathways covering machine learning and artificial intelligence, data, cloud and platform technologies, cybersecurity, architecture, FinOps, quantum computing and software engineering. The offer runs through 31 December 2026.
Eligible users must register with a recognised civil-service or public-sector work email. The government says central and local government, health, education and other public bodies are in scope, while contractors are not. A worker whose legitimate public-sector domain is rejected can contact the programme team. That distinction matters for mixed teams: a civil servant and a contractor doing similar work may not receive the same subsidised access.[1][2]
Free does not mean every course or exam works the same way
Participating providers named by government include AWS, Celonis, Databricks, Datadog, Google, HashiCorp, Microsoft, MongoDB, Oracle, ProsperOps, QA, ServiceNow, Snowflake, the STFC Hartree Centre and WSO2. Some pathways provide learning only, while others include a certification voucher. The launch post says vouchers made available through the campaign must be claimed by 31 December, but some may then remain redeemable for up to six months. Provider-specific rules and deadlines control the actual redemption period.
A sensible learner should begin with a work problem rather than a badge. Someone responsible for procurement may need model-risk, data-governance and evaluation skills; a service owner may need to understand failure modes and human escalation; an engineer may need secure deployment or data-pipeline practice. Checking prerequisites, assessment format, time commitment, accessibility, data handling and whether an exam voucher is actually included can prevent a nominally free pathway from becoming unusable.[1][2]
The 2025 totals show demand, not yet programme impact
Government says the 2025 campaign attracted more than 10,000 sign-ups across 336 public-sector organisations and produced more than 1,200 recognised certifications. Those are useful administrative counts, but the denominator behind the certification figure is unclear. A registration can cover a person, a pathway or an expression of interest, and not every registered learner may have selected a certification-bearing route. Dividing 1,200 by 10,000 would therefore create an unsupported completion rate.
The release does not report how many learners started a course, completed training without certification, attempted an exam, passed, repeated an assessment or used the skill at work. It also does not separate AI pathways from other technology topics, or results by department, seniority, region, sex, disability, ethnicity or prior experience. A testimonial describes one employee's successful move into digital work, but one case cannot establish the programme's average effect.[1][2]
Managers should connect training to supervised practice
A certificate can show that a learner met one provider's assessment standard; it does not by itself demonstrate safe use of AI in a public service. Teams can make the offer more valuable by pairing learning with a bounded assignment: evaluate a model against an approved dataset, write a risk register, compare human and automated decisions, test an accessibility workflow or document when a system must escalate to a person. Work should use authorised tools and non-sensitive data unless governance teams have approved more.
Managers also need to protect time for study. If training is added on top of a full workload, uptake may skew toward people with more autonomy or fewer caring responsibilities. Recording enrolment, starts, completions, assessment results and supervised workplace application would show where support is needed. The dedicated Women in Tech strand may broaden participation, but its effect will depend on who joins, completes and progresses—not only on the number of events or promotional materials.[1][2]
What evidence would show the programme is working
A credible evaluation would publish a pathway-level funnel: eligible visitors, registrations, starts, completions, exam attempts, passes and certifications, with definitions that prevent double counting. It would measure knowledge or practical performance before and after training and follow learners to see whether they changed role, improved service quality, reduced reliance on contractors or avoided material AI and cybersecurity errors. Results should be separated by prior experience and demographic group so apparent success does not hide unequal access.
Confidence would rise if the government publishes those denominators and links training to independently reviewed workplace outcomes. It would weaken if the catalogue produces many registrations but few starts, if vouchers run out without transparent allocation, or if certificates do not translate into safer and more effective public services. For now, the practical conclusion is simple: eligible workers have a time-limited no-fee route into AI and technology learning, while the claim that it builds public-sector capability remains only partly measured.[1][2]