Back to blogCompliance

    Fundamental rights impact assessment: the draft Article 27 template is open for consultation

    The Commission has published the draft questionnaire for the FRIA. Who has to carry out the assessment, what information is needed to fill it in and why to apply for the test by 4 December.

    ZeroFive.AI November 16, 2026Updated on November 19, 2026 9 min

    In brief. On 16 November 2026 the European Commission opened a consultation on the draft template for the fundamental rights impact assessment required by Article 27 of the AI Act, together with a draft explanatory notice. Feedback can be submitted until 15 January 2027, and until 4 December 2026 deployers covered by the obligation can apply to test the template on one of their own systems with support from the AI Office. The obligation applies from 2 December 2027 and concerns public bodies, private entities providing public services, and banks and insurers for certain specific uses. Filling in the template requires information that many organisations have not yet collected.

    After the postponement introduced by the Digital Omnibus, many companies moved the high-risk chapter to the bottom of the agenda, on the assumption that nothing needed doing before December 2027. This week's publication says otherwise. The instrument deployers will use to document the assessment and report its results to the authority already exists in draft, it is open for comment, and those who fall under the obligation can try it in advance.

    The consultation opened on 16 November

    The Commission published two documents, the draft questionnaire template and a draft explanatory notice. The template was prepared by the AI Office under Article 27(5) of the AI Act, with input from the EU Agency for Fundamental Rights, and according to the Commission's announcement it follows the structure of paragraph 1 of the same article, with questions that guide the deployer item by item. The explanatory notice addresses the points on which the text of the regulation leaves most doubt, namely which deployers must carry out the assessment, when it has to be done and how it fits with the obligations of the provider of the system.

    The public questionnaire on EU Survey is open to everyone until 15 January 2027. In parallel, deployers of high-risk systems that will be subject to Article 27 can express interest by 4 December 2026 in testing the draft: those selected apply the template to one of their own systems with assistance from the AI Office, which will choose participants so as to cover different types and sizes of organisation.

    Once it has reviewed the feedback the Commission will adopt the final version, with no date indicated so far. After adoption the template will also be available as an automated tool on the AI Act Single Information Platform, and that will be the format in which deployers prepare the assessment and notify its results to the market surveillance authority.

    Who has to carry out the assessment

    The fundamental rights impact assessment, known as the FRIA, does not concern every deployer of a high-risk system. Article 27 imposes it on three groups, and only for Annex III systems.

    Bodies governed by public law and private entities providing public services must carry out the assessment for all the Annex III systems they use, with the sole exception of those used as safety components in critical infrastructure. The category of private entities providing public services is the least defined: Recital 96 links it to activities such as education, healthcare, social services, housing and the administration of justice, and its boundaries are one of the topics of the explanatory notice.

    For banks and insurers the criterion is the use made of the system, whatever the nature of the organisation using it. The obligation covers anyone who uses systems to evaluate the creditworthiness of natural persons or to establish their credit score, except for systems used to detect financial fraud, and anyone who uses them for risk assessment and pricing in life and health insurance. These are points 5(b) and 5(c) of Annex III, and they are the reason why banks, financial intermediaries and insurance companies are among the most direct addressees of this consultation.

    Being outside these three groups does not mean being outside high risk. A manufacturing company that uses a system to screen candidates or to evaluate employee performance is the deployer of an Annex III system and is bound by the obligations of Article 26, from human oversight to log retention, but does not have to carry out the Article 27 assessment unless it also provides a public service.

    Deployer and use of the systemArticle 27 assessment
    Body governed by public law, Annex III systemsYes, except critical infrastructure
    Private entity providing public services, Annex III systemsYes, except critical infrastructure
    Creditworthiness and credit scoring of natural personsYes, except fraud detection systems
    Risk assessment and pricing in life and health insuranceYes
    Other private company with Annex III systems, for example in recruitmentNo, Article 26 obligations apply
    High-risk systems embedded in regulated products under Annex INo

    The elements of the assessment

    Paragraph 1 of Article 27 lists six elements, and the draft template is built on them.

    PointRequired element
    (a)Description of the deployer's processes in which the system will be used, in line with its intended purpose
    (b)Period of time and frequency with which the system is intended to be used
    (c)Categories of natural persons and groups likely to be affected by its use in the specific context
    (d)Specific risks of harm to those categories, taking into account the information given by the provider under Article 13
    (e)Implementation of human oversight measures, according to the instructions for use
    (f)Measures to be taken if those risks materialise, including internal governance and complaint mechanisms

    The assessment has to be done before first use. In similar cases the deployer may rely on assessments already carried out, its own or the provider's, and when one of the elements changes or is no longer up to date it has to update the information. The results are notified to the market surveillance authority by submitting the completed template, subject to the exemption for the cases covered by Article 46(1).

    The relationship with the data protection impact assessment under the GDPR is governed by paragraph 4. Where a DPIA already covers part of the required elements, the fundamental rights assessment complements it without repeating it, and following the amendments made by Regulation (EU) 2026/1744 the template must allow cross-references to the corresponding sections of the DPIA. For organisations whose record of processing activities is in order this is a practical advantage, because part of the work already exists and only needs to be connected.

    Which framework does your company actually need?

    AI Rating measures maturity across the four areas of the model and shows where to start, with priorities and estimated effort.

    Start your AI Rating

    The timeline and the penalties in Italy

    Article 27 follows the date of the obligations for Annex III systems, which the Digital Omnibus moved from 2 August 2026 to 2 December 2027. From that day a system within scope cannot be put into use without an assessment that has been carried out and notified.

    In Italy the breach already has a defined price. Article 23(4) of Legislative Decree 179/2026 sets a cap of EUR 1 million or 0.5% of annual worldwide turnover for the fundamental rights impact assessment and for the right to an explanation under Article 86, lower than the other bands, as we set out in our analysis of the decree's penalties. The same decree asks the authority to distinguish between substantive and formal obligations, and an assessment that is well filled in but disconnected from how the system is really used risks ending up in the second category.

    The information needed to fill in the template

    The template simplifies the form of the assessment and leaves the work that precedes it untouched. Each item in paragraph 1 presupposes information the organisation must hold before opening the questionnaire, and in our experience that is where the delay builds up.

    Describing the processes requires an inventory of the AI systems in use, with their risk classification and the organisation's role in relation to each one, because a company that has substantially modified a purchased system may find itself in the position of the provider. Assessing the specific risks requires the instructions for use and the Article 13 information, which have to be requested from the provider by contract and do not always arrive complete. For human oversight the organisation must be able to say who checks the outputs, with what powers to intervene and with what training. The last item requires a complaint channel that an affected person can find and use, and someone with the mandate to respond.

    ItemInformation to holdWhere it usually sits
    (a) ProcessesSystems in use, purpose, the organisation's roleAI systems register, process mapping
    (b) Period and frequencyExpected duration, volumes, frequency of useProcess owner, rollout plan
    (c) Affected personsCategories involved, vulnerable groupsDPIA, record of processing activities
    (d) Specific risksInstructions for use, known limits of the systemProvider documentation, contract
    (e) Human oversightWho checks, with what powers and trainingInternal procedures, training plan
    (f) Measures and complaintsEscalation, complaint channel, ownerIncident management, customer service

    These are the same questions an organisation should ask itself before authorising a use case, regardless of the obligation. In the AI Rating model we use with clients they sit in the Risk dimension, together with risk governance and traceability, and a critical gate failed on these points cannot be offset by a high score on technology.

    Voluntary use for those outside the scope

    Companies that do not fall within the three groups of Article 27 can still use the structure of the template as an internal grid. The six items fit any system that affects people, from recruitment to shift allocation to the handling of customer complaints, and adopting them voluntarily produces documentation that is useful for the obligations of Article 26, for the DPIA and for discussions with employee representatives.

    There is also a commercial reason. Banks, insurers and public administrations will ask their suppliers for the information they need to complete their own assessment, and those who sell AI-based systems or services to these organisations will find themselves answering the same questions from the other side of the contract.

    The next deadlines

    DateDeadline
    4 December 2026Deadline to apply for testing the draft with the AI Office
    15 January 2027Public consultation on EU Survey closes
    To be definedAdoption of the final template and release of the automated tool
    2 December 2027Obligation applies to Annex III systems

    For those who fall under the obligation the nearest date is 4 December 2026, the deadline to apply for the test. Taking part means applying the template to a real system with assistance from the AI Office, a year before the obligation applies, and finding out in time which information is missing.

    The public consultation closes on 15 January 2027. It makes sense to respond above all on the points where the draft does not match how decisions are taken inside organisations, for example the level of detail required for processes or the cross-references to the DPIA, because the final template will be the standard against which authorities read the assessments. Then comes 2 December 2027, and the distance between today and that date is the time it takes to build an inventory, fix the contracts with providers and train the people who exercise oversight.

    If you want to understand which of your systems are in scope and which information you are missing to fill in the template, you can start an AI Rating, book an assessment call or write to us at hello@zerofive.ai.

    This article is based on the European Commission's announcement of 16 November 2026 and on the text of Article 27 of the AI Act. The draft template and the explanatory notice are consultation documents and may change before adoption. The article describes the regulatory framework and its organisational implications, and does not constitute legal advice.


    Main sources: European Commission, consultation on the draft template for the fundamental rights impact assessment, 16 November 2026; consultation questionnaire on EU Survey; expression of interest in testing the template; Regulation (EU) 2024/1689, Articles 13, 26, 27, 46 and 86, Recital 96 and Annex III, as amended by Regulation (EU) 2026/1744; Italian Legislative Decree No. 179 of 7 October 2026, Article 23 (Official Gazette, General Series, No. 234 of 8 October 2026).

    Frequently asked questions

    What is the FRIA under the AI Act?

    It is the fundamental rights impact assessment that Article 27 of the AI Act requires from certain deployers before the first use of an Annex III high-risk system. It covers six elements, from the description of processes to the measures to be taken if risks materialise, and its results must be notified to the market surveillance authority.

    Who has to carry out a fundamental rights impact assessment?

    Bodies governed by public law, private entities providing public services, and anyone using systems to evaluate the creditworthiness of natural persons or for risk assessment and pricing in life and health insurance. Other deployers of high-risk systems remain subject to Article 26 but not to Article 27.

    Until when is the consultation on the draft template open?

    The public questionnaire on EU Survey is open until 15 January 2027. Deployers that will be subject to the obligation can also apply by 4 December 2026 to test the draft on one of their own systems with support from the AI Office.

    When does the Article 27 obligation apply?

    From 2 December 2027, the date set by Regulation (EU) 2026/1744 for the obligations on Annex III high-risk systems. The assessment has to be carried out before the first use of the system.

    What penalties apply in Italy for failing to carry out the assessment?

    Article 23(4) of Legislative Decree 179/2026 sets a cap of EUR 1 million or 0.5% of annual worldwide turnover for breaches relating to the fundamental rights impact assessment and the right to an explanation.

    Want to discuss this for your company?

    30 minutes with us to figure out where to start, or an AI Rating to measure your starting point.

    #FRIA#AI Act#Article 27#fundamental rights#high-risk AI#AI Office#AI Rating
    Share

    Keep reading