Activities of Adam JARUBAS related to 2020/2012(INL)
Opinions (1)
OPINION with recommendations to the Commission on a framework of ethical aspects of artificial intelligence, robotics and related technologies
Amendments (24)
Amendment 18 #
Draft opinion
Recital C
Recital C
C. Whereas AI solutions may benefit society in the areas of green transition, environment protection, waste management, circular economy, climate change, energy management and efficiency, air quality e.g. smart grids and electro-mobility;
Amendment 37 #
Draft opinion
Recital E
Recital E
E. Whereas current policy and ethmany ethical challenges related to AI applications and indicated in the Commission White Paper on Artificial guidelines for AIIntelligence, are already addressed by the Union acquis, e.g. risk-assessment processes in place for AI-based health solutions in the Single Market; whereas other areas are lagging behind ethical challenges that must be identified and mitigated, since AI has tremendous capability to threaten patient preference, safety, and privacy; whereas the boundaries between the roles of physicians and machines in patient care need to be outlined;
Amendment 46 #
Draft opinion
Recital F
Recital F
F. Whereas AI solutions may benefit society in the area of food safety, andmong others supporting precision farming or more broadly Farming 2.0, where the Union holds leadership in AI applications, and which will allow to combine more effective production with higher environmental standards and better utilization of resources;
Amendment 64 #
Draft opinion
Paragraph 1 a (new)
Paragraph 1 a (new)
Amendment 73 #
Draft opinion
Paragraph 2 a (new)
Paragraph 2 a (new)
2a. Considers that for all AI applications, developed in the Union and outside of it, the same level of protection must be secured in the Union as it is with all other technologies, including effective judicial redress for parties negatively affected by AI systems, whilst technological innovation needs to be allowed to continue to develop; considers furthermore that this AI risk area is crucial e.g. for the health services, transport involving autonomous vehicles and food safety; calls for a clear distribution of obligations, rights and liabilities among the economic operators involved in AI applications delivery, to attribute each obligation to the actor(s) who is (are) best placed to address any potential risks, whether this is the developer, the deployer, the producer, the distributor or importer, the service provider, the professional or private user, and in this regard for adequate revision of relevant EU legislation, e.g. of the Product Liability Directive and for the harmonization of national legislation; supports the Commission position expressed in the White Paper that due to the complexity of AI systems, securing effective level of protection and redress may require adapting the burden of proof required by national rules on liability for damage caused by the operation of AI applications; is of the opinion that clarity as to legal liability in the AI sector will strengthen enforcement of Union ethical values embodied in its acquis, legal certainty and predictability, and social acceptance supporting the development of a Union AI ecosystem of excellence by pooling investors and increasing market uptake;
Amendment 79 #
Draft opinion
Paragraph 2 b (new)
Paragraph 2 b (new)
2b. Emphasises that AI applications in health should always have the aim of maximising the opportunities they can bring – such as improving the health of individual patients as well as the performance of Member States’ public health systems – without lowering ethical standards and without threatening the privacy or safety of citizens;
Amendment 83 #
Draft opinion
Paragraph 3
Paragraph 3
3. Welcomes the fact that the Risk- Based Approach methodology defined in the Commission White Paper of 19 February 20201 recognises healthcare, transport and energy as high risk sectors by default, and proposes to introduceing listed AI requirements beyond existing-Union rules in these sectors, unless the manner in which AI is used does not involve significant risk; stresses that the Union AI ethical framework should address especially the above high-risk sectors; _________________ 1White Paper On Artificial Intelligence - A European approach to excellence and trust, COM(2020)0065
Amendment 85 #
Draft opinion
Paragraph 3 a (new)
Paragraph 3 a (new)
3a. Calls for clear, objective and transparent procedures at Union level for establishing a public catalogue of AI high-risk applications involving a periodic review and update mechanism; calls for consideration of putting the burden of proof in such procedures, for all AI applications in all domains, on the entity seeking to develop or deploy the AI system, in order to maintain the catalogue open for innovation and avoid ignoring the risk of classifying AI applications as being no high risk;
Amendment 86 #
Draft opinion
Paragraph 3 b (new)
Paragraph 3 b (new)
3b. Recommends supplementing the Risk-Based Approach with an Algorithmic Impact Assessment drawing information for example from the Regulatory Impact Analysis (RIA), GDPR Risk Assessment Procedure, Human Rights Impact Assessment (HRIA) with the results made publicly viewable;
Amendment 89 #
Draft opinion
Paragraph 3 c (new)
Paragraph 3 c (new)
3c. Welcomes the voluntary labelling initiative for non-high risk AI;
Amendment 91 #
Draft opinion
Paragraph 3 d (new)
Paragraph 3 d (new)
3d. Welcomes the Commission commitment expressed in the White Paper to examine safety and liability challenges that are distinctive to healthcare e.g. AI systems providing specialized medical information to physicians or directly to the patient, AI systems performing medical tasks themselves directly on a patient; calls for corresponding examination of the other listed sectors that are by default high-risk ones;
Amendment 93 #
Draft opinion
Paragraph 4
Paragraph 4
4. Notes that, withdue to the fact that legal regulations respond better to current well-defined challenges and due to the rapid development of AI andresulting in the uncertainty that lies ahead, a common, legally well-anchored, Union AI ethical framework willmay expand an ecosystem of trust as defined in the Commission White Paper, whether in environment protection, healthcare or food safety applications, thus supporting the ecosystem of excellence in legal certainty and predictability, providing effective response to the challenges yet not defined in courtrooms, management meetings or scientific laboratories;
Amendment 100 #
Draft opinion
Paragraph 4 a (new)
Paragraph 4 a (new)
4a. Stresses that an law-based Union AI ecosystem of trust, whether regarding the environment protection, health or food safety applications, extended by the Union AI ethical framework, will reinforce legal certainty and predictability, encourage stakeholders' involvement, increase the volume of entrusted data and market up take, allow for economies of scale and support an ecosystem of excellence in those sectors; is of the opinion that this will strengthen the Union AI sector's global competitiveness and the potential to promote Union values and standards;
Amendment 107 #
Draft opinion
Paragraph 5
Paragraph 5
5. Strongly supports the Commission in establishing a common Union AI ethical framework to counter the shortcomings caused by AI internal market fragmentation, including environmental, healthcare, and food safety applications, and to prevent AI double standards across Member States for AI developed in Union and beyond, inter alia in areas such as consumer data management, protection and privacy in smart grids, waste management, equal access to services, patient-doctor relationship standards, data protection and privacy, civil liability in AI-assisted public healthcare, civil liability regarding autonomous vehicles or machinery; calls for proper legal anchoring and positioning of such a Union AI ethical framework;
Amendment 114 #
Draft opinion
Paragraph 5 a (new)
Paragraph 5 a (new)
5a. Underlines the importance of training highly skilled professionals in environmental, health and food safety- related areas that should go hand-in-hand with the need of having diverse teams of developers and engineers working alongside key actors to prevent gender and cultural bias of being inadvertently included in AI algorithms, systems and applications. The mutual recognition of such qualifications across the Union should be also ensured;
Amendment 130 #
Draft opinion
Paragraph 6 a (new)
Paragraph 6 a (new)
6a. Stresses that the issue of opacity of the AI “black box-effect” machine deep learning, its complexity and partially autonomous behaviour might prevent the enforcement of the Union acquis, including its ethical values and standards and undermine the Union AI ecosystem of trust among investors and consumers, and thus hamper the AI ecosystem of excellence of the Union;
Amendment 131 #
Draft opinion
Paragraph 6 a (new)
Paragraph 6 a (new)
6a. Calls on the Commission to promote and fund the development of human-centric artificial intelligence, robotics and related technologies that address environment and climate challenges and that ensure equal access to and enjoyment of fundamental rights through the use of tax, green public procurement, or other incentives;
Amendment 135 #
Draft opinion
Paragraph 6 b (new)
Paragraph 6 b (new)
6b. Calls for Union guiding initiatives promoting interpretable algorithms, eXplainable AI (xAI), symbolic reasoning AI, white box AI testing technics, by showing that those technologies can be combined with deep neural networks and by showing its legal, ethical and often business advantages, and also promoting methods to determine risks connected with different technological options using among others the experience of the UK’s Information Commissioner's Office (ICO) and The Alan Turing Institute guidelines “Explaining decisions made with AI”, showing that even highly complex neural AI systems can be interpreted sufficiently;
Amendment 137 #
Draft opinion
Paragraph 6 c (new)
Paragraph 6 c (new)
6c. Welcomes a European strategy for data, addressing challenges ahead for the Union in this area that is key to AI progress, and seeking European opportunities for competitive advantage in new data economy, especially in the growing sector of decentralised, non- personal data coming from industry, business and the public sector and from devices at the edge of the network, which is expected to constitute 80% of 175 zettabytes in 2025 and reverse current proportions;
Amendment 138 #
Draft opinion
Paragraph 6 d (new)
Paragraph 6 d (new)
6d. Stresses that a data economy fuelling AI development, which is called therefore the crude oil of 21st century, is not without risks; underlines that AI ethical and thus legal challenges extend beyond data protection and privacy, to liability for data environment used for AI training, testing and, in some cases, to data input used for algorithm self- adaptation during functioning, which may result in bias, discrimination or mistakes; welcomes requirements proposed in the White Paper for high risk AI training data, addressing as well safety – sufficiently broad data to cover all relevant scenarios in order to avoid dangerous situations as discrimination - sufficiently representative data to reflect well the social environment it will be applied to;
Amendment 141 #
Draft opinion
Paragraph 7
Paragraph 7
7. Supports the view that the seven AI requirements identified in the Ethics Guidelines for Trustworthy AI of the High- Level Expert Group on AI constitute solid building blocks for a common Union AI ethical framework, with proper legal anchoring, addressing, among others, ethical aspects of AI applications in environment, health and food protection; calls for an improvement of the acquis on transparency, traceability and human oversight, which were indicated as areas in need of further improvement in the feedback given on the Guidelines by 350 organisations; furthermore, encourages the creation of the Union AI ethical framework in a spirit of openness to the works of other international partners that share Union values, e.g. to the Rome Call for AI Ethics by Pope Francis;
Amendment 151 #
Draft opinion
Paragraph 8 a (new)
Paragraph 8 a (new)
8a. Stresses that the development of AI applications might bring down the costs and increase the volume of services available, e.g. health services, public transport, Farming 2.0, making them more affordable to a wider spectrum of society; stresses that AI applications may also result in the rise of unemployment, pressure on social care systems, increase of poverty; emphasizes in accordance with the values enshrined in Article 3 of the Treaty on European Union the need to adapt Union AI transformation to socio- economic capacities, adequate social shielding, education and creation of alternative jobs; calls to consider the establishment of a Union AI Adjustment Fund building upon the experience of The European Globalisation Adjustment Fund (EGF) or the currently developed Just Transition Fund;
Amendment 158 #
Draft opinion
Paragraph 8 b (new)
Paragraph 8 b (new)
8b. Calls for the education of professionals in the area of environment protection, health and food safety preparing for AI applications and rising awareness of AI risks and ethical challenges;
Amendment 160 #
Draft opinion
Paragraph 8 c (new)
Paragraph 8 c (new)