# HD-INC-223: Italy's Top Court Rejected a Ministry Reading That Let Teacher Postings Skip Higher Ranks > Italy's Court of Cassation held that the Education Ministry's reading of its own rules, applied through a fully computerised procedure for assigning supply-teaching posts, was unreasonable because it could let a post go to a lower-ranked candidate. The court set the binding interpretation on a referral from an appeal court, where the ministry's appeal in the original case is still pending. - Source: https://useheadlights.com/library/HD-INC-223/ - Publisher: Headlights (useheadlights.com), author Ellie Harris - Published: 2026-10-03 - Company: Ministry of Education and Merit (Ministero dell'istruzione e del merito) - Sector: Employment services - Country: Italy - Year: 2026 - Behaviour in the record (Headlights reading): Not established - Governance domains: Administrative law, Algorithmic accountability - Licence: CC BY-NC 4.0. Cite as: Harris, E. (2026). Italy's Top Court Rejected a Ministry Reading That Let Teacher Postings Skip Higher Ranks. Headlights AI Incident Library, HD-INC-223. https://useheadlights.com/library/HD-INC-223/ ## What happened It was reported that teachers on Italy's provincial ranking lists apply for year-long supply posts by listing the schools they want. The Court of Cassation, Labour Section, records in judgment n. 18156/2026, published on 5 June 2026, that under ministerial order n. 112 of 6 May 2022 the assignment procedure was fully computerised. It was reported that a teacher in the province of Chieti received no post for 2022/2023. Schools she had chosen became free in a later round and were given to candidates with lower scores than hers. She challenged what the court describes as the computerised procedure for assigning annual supply posts. The judgment records that the first-instance court partly upheld her claim, found the procedure unlawful and awarded damages. It was reported that the ministry appealed to the L'Aquila Court of Appeal, arguing that its approach served the economy and efficiency of the procedure. ## What an auditable version would have shown It was reported that lower courts were divided. The judgment records that some found the procedure unlawful, while several appeal courts found it reasonable and consistent with efficiency. The court records that the L'Aquila Court of Appeal referred the question to the Court of Cassation on 17 October 2025 because the same dispute could arise elsewhere. It was reported that the Court of Cassation rejected the ministry's interpretation of the order. The court held that the interpretation had no sure basis in the text and was not free of intrinsic unreasonableness. The court found that this reading could allow a higher-ranked candidate to be passed over for a lower-ranked candidate. ## Where the gap was It was reported that the same mechanism appeared in almost identical terms in order n. 88 of 16 May 2024. The court records that a new order of 16 February 2026 changed the wording in line with the court's reading. The judgment gives no count of teachers who could have been passed over. It was reported that the judgment does not describe how the ministry checked its reading of the rule before applying it through the computerised procedure. A ConductRecord keeps, for each assignment, which candidates were considered, which were skipped and the rule that skipped them. A MetricRecord is designed to produce signed counts, such as how often a post went to a lower-ranked candidate who chose the same school, that a ministry, court or union can verify. ConductRecord and MetricRecord are Headlights designs. ## What governance should have looked like It was reported that the Court of Cassation stated a principle of law rather than deciding the ministry's appeal. The court held that leaving a school off the list meant a teacher waived only the schools they had not chosen. It also held that if a chosen school became available later, the teacher's position in the ranking still had to be respected. The judgment records that this principle binds the L'Aquila Court of Appeal when it decides the ministry's appeal. It was reported that no decision by L'Aquila had been found as of 3 October 2026. Where a public body builds its reading of a rule into an automated procedure, best practice would be to test that reading against the law, record each candidate skipped and why, and review the results before the next round. **Failure Pattern:** according to the Court of Cassation, the ministry's computerised assignment of supply-teaching posts applied an unreasonable reading of its own order, under which a higher-ranked teacher could be passed over in favour of a lower-ranked one. **Governance Principle:** a rule built into an automated public procedure should be tested against the law it implements before it is applied to everyone. The reference implementation of ConductRecord and MetricRecord is open source. It lives at [github.com/saffronandindia/headlights-oss](https://github.com/saffronandindia/headlights-oss), Apache 2.0 licensed and free to install. The repository is public now. ## Sources - [Corte di Cassazione, Sezione Lavoro, sentenza n. 18156/2026, RG 21065/2025 (published 5 June 2026)](https://www.cortedicassazione.it/resources/cms/documents/18156_06_2026_civ_noindex.pdf) ## Related - [HD-INC-050](/library/HD-INC-050)