lecture 5 Flashcards
Definitions – General-Purpose AI Model
(Art 3 (63) AI Act)
“‘general-purpose AI model’ means an AI model, including where such an AI model is
trained with a large amount of data using self-supervision at scale, that displays
significant generality and is capable of competently performing a wide range of
distinct tasks regardless of the way the model is placed on the market and that can be
integrated into a variety of downstream systems or applications, except AI models that
are used for research, development or prototyping activities before they are released
on the market”
Definitions – Operator
(Art 3 (8) AI Act)
“operator means a provider, product manufacturer, deployer, authorised representative, importer or distributor”
Definitions – Provider & User
Providers (Art 3 (3) AI Act)
* ‘provider’ means a natural or legal person, public authority, agency or other body that develops an AI system or a
general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the
market or puts the AI system into service under its own name or trademark, whether for payment or free of charge
Deployer (Art 3 (4) AI Act)
* ‘deployer’ means a natural or legal person, public authority, agency or other body using an AI system under its
authority except where the AI system is used in the course of a personal non-professional activity
Definitions – Authorised Representative & Importer & Distributor
Authorised representative (Art 3 (5) AI Act)
* ‘authorised representative’ means a natural or legal person located or established in the Union who has received
and accepted a written mandate from a provider of an AI system or a general-purpose AI model to, respectively,
perform and carry out on its behalf the obligations and procedures established by this Regulation
Importer (Art 3 (6) AI Act)
* ‘importer’ means a natural or legal person located or established in the Union that places on the market an AI
system that bears the name or trademark of a natural or legal person established in a third country
Distributor (Art 3 (7) AI Act)
* ‘distributor’ means a natural or legal person in the supply chain, other than the provider or the importer, that
makes an AI system available on the Union market
AI Act
f) Scope of Application
- providers (Art 2 (1) lit a AI act)
- deployers (Art 2 (19 lit AI Act)
- providers and deployers (Art 2 (2) lit c AI Act)
- Importers & distributors (Art 2 (1) lit d AI Act)
- Product manufacturers (Art 2 (1) lit e AI Act)
- Authorised representatives (Art 2 (1) lit f AI Act)
- Affected persons (Art 2 (1) lit g AI Act)
Scope of Application (Exceptions)
- Military, defence, national security (Art 2 (3) AI Act)
- Scientific research and development (Art 2 (6) AI Act)
- Research, testing or development of AI systems and models prior to their
being placed on the market or put into service (Art 2 (8) AI Act) - Personal non-professional activities (Art 2 (10) AI Act)
AI Act
g) Risk Categories
- Prohibited AI Practices
(Art 5 AI Act)
AI systems that deploy subliminal techniques and are likely to cause significant harm (Art 5 (1)
lit a AI Act)
Exploiting AI systems, which target vulnerable groups and are likely to cause significant harm (Art 5 (1) lit b AI
Act)
Social scoring AI systems (Art 5 (1) lit c AI Act)
Criminal risk prediction (Art 5 (1) lit d AI Act)
Biometric Systems (Art 5 (1) lit e-h AI Act)
High Risk AI Systems
(Art 6 in connection with Annex I/III AI Act )
1) Regulated products or safety components of regulated products
(e.g. medical devices, machinery) which are subject to third-party conformity assessment
under the relevant EU legislation (→ Art 6 (1) in connection with Annex I)
2) Certain AI systems in the areas listed below
unless the system does not pose a significant risk of harm to the health, safety or fundamental rights of natural persons,
including by not materially influencing the outcome of decision making (→ Art 6 (2) in connection with Annex III):
- Biometrics
✓ Critical infrastructure
✓ Education and vocational training
✓ Employment, workers management and access to
self-employment
✓ Access to and enjoyment of essential private
services and essential public services and benefits
✓ Law enforcement
✓ Migration, asylum and border control management
✓ Administration of justice and democratic processes
High Risk AI Systems – Requirements
(Chapter III, Section 2)
Establish and implement a risk management system (Art 9 AI Act)
Use high-quality training, validation and testing data (relevant, representative etc) (Art 10 AI Act)
Draw up documentation (Art 11 AI Act) & record-keeping (Art 12 AI Act)
Ensure appropriate degree of transparency and provide deployers with information (on how to use the system, its capabilities and limitations)
(Art 13 AI Act)
Human oversight (Art 14 AI Act)
Accuracy, robustness and cybersecurity (Art 15 AI Act)
High Risk AI Systems – Obligations
(Chapter III, Section 3)
provider obligations:
► Ensure compliance of high-risk AI systems with the requirements of the AI Act
► Provide contact information on the high-risk AI system or on the packaging or its accompanying documentation
► Establish and implement quality management system
► Keep up-to-date documentation
► Ensure conformity assessment
► Register AI system (listed in Annex III) in EU database
► Conduct post-market monitoring and take corrective actions
► Report serious incidents
► Collaborate with market surveillance authorities
deployer obligations:
► Operate high-risk AI system in accordance with instructions of use
► Ensure human oversight & monitor operation for possible risks
► Keep automatically generated logs
► Inform about any serious incidents
► Carry out fundamental rights impact assessment (deployers = bodies governed by public law, or private entities providing public serv.)
Limited Risk AI Systems
(Art 50 AI Act)
Systems interacting directly with natural persons (Art 50 (1) AI Act)
Systems generating synthetic audio, image, video or text content (Art 50 (2) AI Act)
Emotion recognition or biometric categorisation systems (Art 50 (3) AI Act)
Deep fake AI systems (Art 50 (4) AI Act)
Limited Risk AI Systems – Transparency Obligations
provider obligations:
► Systems interacting with natural persons:
► ensure that AI systems intended to interact with natural persons inform natural persons that they are interacting with an AI system
► unless this is obvious from the point of view of a reasonably well-informed, observant and circumspect person
► Systems generating synthetic data
► ensure that the outputs of the AI system are marked in a machine-readable format and detectable as artificially generated or manipulated
deployer obligations:
► Biometric categorisation & emotion recognition systems: inform the exposed natural persons about
the system’s operation
► Deep fake systems: disclose that the content has been artificially generated or manipulated
!!!!! Exemptions: e.g.
criminal offences/
investigations!!!
General Purpose AI Models
(Art 51-56 AI Act)
Art 51 AI Act: Classification of general-purpose AI models as general-purpose AI models with systemic risk
Art 52 AI Act: Procedure
Art 53 AI Act: Obligations for providers of general-purpose AI models
Art 54 AI Act: Authorised representatives of providers of general-purpose AI models
Art 55 AI Act: Obligations for providers of general-purpose AI models with systemic risk
Art 56 AI Act: Codes of practice
