Article 9

(k) processing in the context of the development and operation of an AI system as defined in Article 3, point (1), of Regulation (EU) 2024/1689 or an AI model, subject to the conditions referred to in paragraph 5.

(k) processing in the context of the development and operation of an AI system as defined in Article 3, point (1), of Regulation (EU) 2024/1689 or an AI model, subject to the conditions referred to in paragraph 5.

(k) processing in the context of the development and operation of an AI system as defined in Article 3, point (1), of Regulation (EU) 2024/1689 or an AI model, subject to the conditions referred to in paragraph 5.

Justification

Specific categories of data shall not become the general data base for AI. A subsequent removal, which may be waived in the event of ‘disproportionate effort’, does not constitute effective protection.

(k) processing in the context of the development and operation of an AI system as defined in Article 3, point (1), of Regulation (EU) 2024/1689 or an AI model, subject to the conditions referred to in paragraph 5.

(k) processing in the context of the development and operation of an AI system as defined in Article 3, point (1), of Regulation (EU) 2024/1689 or an AI model, subject to the conditions referred to in paragraph 5.

(k) processing in the context of the development and operation of an AI system as defined in Article 3, point (1), of Regulation (EU) 2024/1689 or an AI model, subject to the conditions referred to in paragraph 5.

(k) processing in the context of the development and operation of an innovative system or technology that is necessary for the legitimate interests pursued by the controller or by a third party, whether or not such system or technology exists at the time of adoption of this Regulation, of which an AI system as defined in Article 3, point (1), of Regulation (EU) 2024/1689 or an AI model is a current example, subject to the conditions referred to in paragraph 5, and without prejudice to the prohibited practices laid down in Article 5 of Regulation (EU) 2024/1689, the classification rules for high-risk AI systems laid down in Article 6, the obligations of deployers of high-risk AI systems laid down in Article 26, and the transparency obligations laid down in Article 50 of that Regulation.

(k) incidental and residual processing in the context of the development, training, testing, deployment and operation of an AI system as defined in Article 3, point (1), of Regulation (EU) 2024/1689 or an AI model as referred to in Regulation (EU) 2024/1689, including for the improvement of such systems or models and of related products or services, subject to the conditions referred to in paragraph 5.

(k) strictly incidental and residual processing in the context of the development and operation of an AI system as defined in Article 3, point (1), of Regulation (EU) 2024/1689 or an AI model, subject to the conditions referred to in paragraph 5.

(k) processing in the context of the development and operation of an AI system as defined in Article 3, point (1), of Regulation (EU) 2024/1689 or ana AIGPAI model, subject to the conditions referred to in paragraph 5.

Justification

The amendment clarifies that Article 9 applies where data is processed to reveal sensitive attributes, not where such attributes could merely be inferred incidentally from lawful processing. This reduces uncertainty for AI, safety, mobility, cybersecurity and other public-benefit technologies using large datasets, while preserving strong safeguards. Voluntary biometric use remains allowed only with a comparable non-biometric alternative, and safeguards such as minimisation, access limits and pseudonymisation.

(l) processing of biometric data is necessary for the purpose of confirming the identity of a data subject (verification), where the biometric data or the means needed for the verification is under the sole control of the data subject.’

Justification

In practice, alternative means of identification are almost always available, so the use of biometric data is by no means ‘necessary’. Establishing the ‘necessity’ of the use would also normalise the use of biometric data, which should be avoided.

(l) processing of biometric data is necessary for the purpose of confirming the identity of a data subject (verification), where the biometric data or the means needed for the verification is under the sole control of the data subject.

(l) processing of biometric data is necessary for the purpose of confirming the identity of a data subject (verification), where the biometric data or the means needed for the verification is under the sole control of the data subject.processing of biometric data is necessary for the purpose of confirming the identity of a data subject (one-to-one verification), and the following conditions are met: (i) the identity confirmation is required by Union or Member Statea national law providing suitable and specific measures to safeguard the fundamental rights and the interests of the data subject; (ii) the biometric data and the means needed for the verification are under the sole control of the data subject; (iii) the controller can demonstrate that there are no less intrusive alternative solutions that could achieve the same objective as effectively; and (iv) the biometric data is deleted after the verification process has been completed.

(l) processing of biometric data is necessaryauthorised forwhen theit purposeis ofused confirmingto confirm the identity of a data subject (verification), whereprovided that the biometric data or the means needed for thethat verification isremain under the sole control of the data subject. Where biometric verification is part of an access control process within strategic public transport infrastructure, the storage of biometric data and the means necessary for verification may be under the exclusive control of the controller, provided that such processing is strictly necessary and proportionate, ensures a high level of security, including robust encryption, strict access controls and protection against cyberattacks, and fully respects the fundamental rights and freedoms of the data subjects concerned.

(l) processing of biometric data is necessary for the purpose of confirming the identity of a data subject (verification), where the biometric data or the means needed for the ‘one-to-one’ verification is under the sole control of the data subject, subject to appropriate safeguards to protect the fundamental rights and the interests of the data subject, and provided that the processing is in accordance with this Regulation and does not constitute a new derogation from the prohibition provided for in Article 9(1).’

(l) processing of biometric data is necessary for the purpose of confirming the identity of a data subject (verification), where the biometric data or the means needed for the verification is under the sole control of the data subject.processing of biometric data is necessary for the purpose of verifying the identity or identifying a data subject, provided that biometric data is optional, the data subject has access to a non-biometric alternative of comparable effectiveness and that no photo or video is captured from individuals who do not consent to the facial recognition, and that appropriate safeguards laid down under Union law are applied to ensure the protection of fundamental rights of the data subject.

(l) processing of biometric data is strictly necessary for the purpose of confirming the identity of a data subject (using one-to-one verification), where the biometric data or the means needed for the verificationrecognition is under the sole control of the data subject and occurs on the data subject’s device, except in highly exceptional cases where this is strictly necessary, proportional and uses state of the art privacy preserving technology, such as encryption and zero knowledge proofs.

(l) processing of biometric data is strictly necessary and proportionate for the purpose of confirming the identity of a data subject (verification), where the biometric data or the means needed for the verification is under the sole control of the data subject and subject to appropriate safeguards to protect the fundamental rights and the interests of the data subject, as laid down in Union law or Member State law, in accordance with paragraph 4 of this article.

(l) processing of biometric data is necessary for the purpose of confirming the identity of a data subject (verification), where the biometric data or the means needed for the verification is under the sole control of the data subject.processing of biometric data for the purposes of uniquely identifying a natural person provided that the use of biometric data is not mandatory, that a non-biometric alternative of comparable effectiveness is available to the data subject and that appropriate safeguards are applied to protect the fundamental rights and interests of the data subject.

Justification

The amendment clarifies that Article 9 applies where data is processed to reveal sensitive attributes, not where such attributes could merely be inferred incidentally from lawful processing. This reduces uncertainty for AI, safety, mobility, cybersecurity and other public-benefit technologies using large datasets, while preserving strong safeguards. Voluntary biometric use remains allowed only with a comparable non-biometric alternative, and safeguards such as minimisation, access limits and pseudonymisation.

(l) processing of biometric data is necessary for the purpose of confirming the identity of a data subject (verification), where the biometric data or the means needed for the verification is under the sole control of the data subject and the biometric data is neither centrally stored nor further processed for any other purpose.’

(l) processing of biometric data is necessary for the purpose of confirming the identity of a data subject (verification), where the biometric data or the means needed for the one-to-one verification is under the sole control of the data subject.

Justification

The amendment clarifies that Article 9 applies where data is processed to reveal sensitive attributes, not where such attributes could merely be inferred incidentally from lawful processing. This reduces uncertainty for AI, safety, mobility, cybersecurity and other public-benefit technologies using large datasets, while preserving strong safeguards. Voluntary biometric use remains allowed only with a comparable non-biometric alternative, and safeguards such as minimisation, access limits and pseudonymisation.

(la) In Article 9, paragraph 2, the following point is inserted

(la) In Article 9, paragraph 2, the following point is inserted. the other points should be adapted accordingly.

(lb) In Article 9, paragraph 2, the following point is added

(lb) in Article 9, paragraph 2, the following point is added

(aa) Paragraph 1 of Article 9 is replaced by the following

(aa) In Article 9, the following paragraph is inserted:

(b) the following paragraph shall be added: For processing referred to in point (k) of paragraph 2, appropriate organisational and technical measures shall be implemented to avoid the collection and otherwise processing of special categories of personal data. Where, despite the implementation of such measures, the controller identifies special categories of personal data in the datasets used for training, testing or validation or in the AI system or AI model, the controller shall remove such data. If removal of those data requires disproportionate effort, the controller shall in any event effectively protect without undue delay such data from being used to produce outputs, from being disclosed or otherwise made available to third parties.’

Justification

Specific categories of data shall not become the general data base for AI. A subsequent removal, which may be waived in the event of ‘disproportionate effort’, does not constitute effective protection.

(b) the following paragraph is added: 5. For processing referred to in point (k) of paragraph 2, appropriate organisational and technical measures shall be implemented to avoid v the collection and otherwise processing of special categories of personal data. Where, despite the implementation of such measures, the controller identifies special categories of personal data in the datasets used for training, testing or validation or in the AI system or AI model, the controller shall remove such data. If removal of those data requires disproportionate effort, the controller shall in any event effectively protect without undue delay such data from being used to produce outputs, from being disclosed or otherwise made available to third parties.’

(5) For processing referred to in point (k) of paragraph 2, appropriate organisational and technical measures shall be implemented to avoid the collection and otherwise processing of special categories of personal data. Where, despite the implementation of such measures, the controller identifies special categories of personal data in the datasets used for training, testing or validation or in the AI system or AI model, the controller shall remove such data. If removal of those data requires disproportionate effort, the controller shall in any event effectively protect without undue delay such data from being used to produce outputs, from being disclosed or otherwise made available to third parties.’

5. For processing referred to in point (k) of paragraph 2, appropriate organisational and technical measures shall be implemented to avoid v the collection and otherwise processing of special categories of personal data. Where, despite the implementation of such measures, the controller identifies special categories of personal data in the datasets used for training, testing or validation or in the AI system or AI model, the controller shall remove such data. If removal of those data requires disproportionate effort, the controller shall in any event effectively protect without undue delay such data from being used to produce outputs, from being disclosed or otherwise made available to third parties.

5. For processing referred to in point (k) of paragraph 2, appropriatestrict organisational and technical measures shall be implemented to avoid vprevent the collection and otherwise processing of special categories of personal data. Where, despite the implementation of such measures, the controller identifies special categories of personal data in the datasets used for training, testing or validation, orthe incontroller shall erase such data without undue delay prior to the completion of the training phase of the AI system or AI model. If erasure of those data from an already trained AI system or AI model proves to be technically impossible, the controller shall removefully such data. If removal of those data requires disproportionate effort,document the controllertechnological shallimpossibility, notify the competent supervisory authority, and without undue delay, and in any event, effectively protect withoutsuch unduedata delayby applying state-of-the-art filtering or aligment techniques to guarantee such data fromcannot beingbe further processed, used to produce outputs, from beingre-identified, disclosed or otherwise made available to third parties. Processing referred to in point k of paragraph 2 should exlude the processing of special categories of personal data originating from data provided by, or generated through, the activity of end users on online platforms or core platfrom services as defined in Regulation (EU) 2022/1925, irrespective of whether such processing occurs during the development, training, or operational phase of any AI system or AI model.

5. For processing referred to in point (k) of paragraph 2, appropriate organisational and technical measures shall be implemented to avoid v the collection and otherwise processing of special categories of personal data, without undue delay. In assessing whether the removal of those data requires a disproportionate effort, account shall be taken of the state of the art, the costs of implementation, the volume of data concerned, the technical feasibility of isolating or extracting the data, and the nature, scope and purposes of the processing. Where, despite the implementation of such measures, the controller identifies special categories of personal data in the datasets used for training, testing or validation or in the AI system or AI model, the controller shall remove such data. If removal of those data requiresis disproportionatetechnically effortimpossible despite documented state-of-the-art measures, the controller shall in any event effectively protect without undue delay such data from being used to produce outputs, or re-used for AI outputs, profiling, advertising or other purposes, and from being disclosed or otherwise made available to third parties. In assessing whether the removal of those data requires a disproportionate effort, account shall be taken of the state of the art, the costs of implementation, the volume of data concerned, the technical feasibility of isolating or extracting the data, and the nature, scope and purposes of the processing; or re-used for AI outputs, profiling, advertising or other purposes,

Justification

A general AI derogation should not create a route for the incidental use of patient or health-related data. Health data require a specific legal basis, stronger safeguards and effective supervisory oversight. The same carve-out should be reflected in proposed Article 88c on legitimate interests for AI processing.