Updated 2 November 2026: shortened and corrected against the Action Plan itself, which we now link. We found no separately announced programme: the duty is one sentence in Action 7 of the Plan, led by the Cyber Security Presidency, while the curriculum and micro-certificates come from Action 1, which covers many groups. The September press report combined the two. Unsourced comparisons and general commentary have been removed.

Correction (19 October 2026): We removed references to two earlier reports (on a KVKK fine for AI predictive profiling and on Council of State rulings on algorithmic errors) that have been withdrawn because they could not be verified, and updated the description of Article 4 of the EU AI Act, which the "Digital Omnibus on AI" amended in July 2026.

Türkiye's AI Action Plan (2026-2030) says basic AI literacy training for public personnel will be made mandatory through institutions' human-resources systems. The Plan took effect through Presidential Circular 2026/9, published on 18 August 2026. Press reports on 2 September 2026 brought the point to wider attention. We found no implementing rule or start date as of 1 November 2026.

What does the AI Action Plan say about public servants?

The key sentence is in Action 7 of the Plan: basic AI literacy e-learning and short role-based modules for public personnel "will be made mandatory through human-resources systems", and completion and success rates will be monitored regularly. Action 7 is led by the Cyber Security Presidency, working with the Ministry of Industry and Technology, the Ministry of Treasury and Finance and the Presidency of Strategy and Budget.

The content comes from a different place. Action 1, led by the Ministry of Industry and Technology, sets up a National AI Literacy and Safe Use Programme for many groups, with public personnel among them. It covers:

  • basic AI concepts;
  • safe use of generative AI tools;
  • personal data protection;
  • awareness of misinformation and deepfakes;
  • principles of ethical use.

Those who complete Action 1's programme receive level-based micro-certificates that record skills shown by assessment, not mere attendance.

The report on Memurlar.Net, credited to Anadolu Agency, presented these together. It was a report on the Plan, not a new announcement. Action 7 also asks institutions to name a senior AI manager and to prepare a three-year roadmap within 12 weeks of that appointment. Our note on the circular and our review of the Plan give the wider picture.

Is there a legal basis for mandatory training?

For civil servants, yes, in general terms. Article 214 of Civil Servants Law No. 657 provides for in-service training under regulations, and Article 215 requires a training unit in each institution. The Plan itself is put into effect by a circular addressed to the administration. In our reading, each institution will still need to turn the sentence into its own rules, and more so for contract staff and workers under other regimes. The sentence does not define "public personnel", and the sources we read do not say whether municipalities or state enterprises are covered, or what happens to someone who does not complete the training.

Why does AI literacy matter legally?

Because the risks the training addresses already carry legal consequences. These are our observations, not statements in the Plan.

  • Personal data. Public institutions are controllers under Law No. 6698, so Article 12 on data security applies, and Article 9 may apply if a tool sends data abroad. KVKK's paper on generative AI in the workplace (February 2026) discusses "shadow AI", meaning use outside institutional control. See also our guide to KVKK and AI governance.
  • Secrecy. Article 31 of Law No. 657 bars civil servants from disclosing confidential information without authorisation. Articles 258 and 136 of the Penal Code penalise disclosing official secrets and unlawfully disclosing personal data. Whether a particular upload meets those offences depends on the facts.
  • Decisions. Under Article 2 of Law No. 2577, administrative courts review acts for authority, form, reason, subject and purpose. An official who adopts an AI output still answers for those elements. Article 11 of Law No. 6698 also gives a right to object to an adverse result produced exclusively by automated analysis, subject to the exceptions in Article 28(2).
  • Records. Training scores and certificate levels relate to identifiable staff, so they are personal data. Nothing we read says how they will affect promotion or assignment.

Our guide to an AI use policy covers the written rules that would accompany such training.

How does this compare with the EU?

Article 4 of the EU AI Act is the nearest parallel, but it is softer and does not apply to Turkish public bodies as such. Since the Digital Omnibus entered into force on 27 July 2026, providers and deployers must take measures to support AI literacy among their staff, taking account of knowledge, experience and context. The text says this does not require them to guarantee any specific level. The Turkish sentence goes the other way: mandatory training with monitored results, though it applies only to the public sector.

Frequently asked questions

Is AI training already mandatory for Turkish public servants?

The Türkiye AI Action Plan says it will be made mandatory through human-resources systems. We found no implementing rule or start date as of 1 November 2026, so institutions may not yet be applying it.

Who is responsible for the training?

The Action Plan puts the public-sector action (Action 7) under the Cyber Security Presidency, working with other bodies. The wider national literacy programme (Action 1) is led by the Ministry of Industry and Technology. Press reports described the Ministry as coordinating.

Will there be certificates?

Action 1 of the Action Plan provides level-based micro-certificates based on assessment. Press reports link them to the training of public staff. How they will affect appointment or promotion is not stated in the sources we read.

Burhan Doğuş Ayparlar's View

This section reflects my personal assessment as the founder of this site and an attorney at law (Türkiye).

In my view, mandatory training that measures results rather than attendance is a sensible design. As a lawyer, however, I must stress one point: literacy is the start of a governance system, not the system itself. Staff also need written rules on which tools they may use with which data, and a channel for reporting incidents.

Otherwise training risks shifting responsibility from the institution to the individual. "You had the training, so the mistake is yours" must not be allowed to cover systemic gaps.

Sources: Türkiye AI Action Plan (2026-2030), Actions 1 and 7, as published by the Presidency of Strategy and Budget; Presidential Circular 2026/9 (Official Gazette No. 33344, 18 August 2026); Memurlar.Net, credited to Anadolu Agency, and Dünya (2 September 2026), press reports; Laws Nos. 657, 6698, 5237 and 2577; KVKK paper (February 2026); Regulations (EU) 2024/1689 and 2026/1744. All read on 1 November 2026. Implementing details were not available. General information, not legal advice.