Activities of Sven MIKSER related to 2020/2012(INL)
Shadow opinions (1)
OPINION with recommendations to the Commission on framework of ethical aspects of artificial intelligence, robotics and related technologies
Amendments (30)
Amendment 2 #
Draft opinion
Paragraph 1
Paragraph 1
1. Highlights that the security and defence policies of the European Union and its Member States are guided by the principles enshrined in the European Charter of Fundamental Rights and those of the United Nations Charter, and by a common understanding of the universal values of the inviolable and inalienable rights of the human person, of freedom, of democracy, of equality and of the rule of law; highlights that all defence- related efforts within the Union framework must respect these universal values whilst promoting peace, security and progress in Europe and in the world;
Amendment 7 #
Draft opinion
Paragraph 2
Paragraph 2
2. Stresses that a Union framework regulating the use of artificial intelligence (AI)-enabled systems in defence must respect all applicable legal regimes, in particular the international humanitarian law and the international human rights law, and be in compliance with Union law, principles and values; stresses that the EU should play a global role in leading the way towards a credible and binding AI regulatory agenda rooted in democratic values; calls on the Union to assess the inherent AI-related risks with regard to the application of Union law, and foresee necessary adjustment and enforcement where needed; underlines that emerging technologies not covered by international law should be judged by the principle of humanity and the dictates of public conscience; underlines that the ethics of AI-enabled systems in defence must be assessed from the point of view of Human rights, and notably human safety, health and security, freedom, privacy, integrity and dignity;
Amendment 9 #
Draft opinion
Paragraph 2
Paragraph 2
2. Stresses that a Union framework regulating the use of artificial intelligence (AI)-enabled systems in defence must respect all applicable legal regimes, in particular the international humanitarian law and the international human rights law, and be in compliance with Union law, principles and values; the framework must therefore indicate the likeliness of errors and inaccuracies to deployers for the deployment of AI technology; calls on the Union to assess the inherent AI-related risks with regard to the application of Union law, and foresee necessary adjustment and enforcement where needed;
Amendment 14 #
Draft opinion
Paragraph 2 a (new)
Paragraph 2 a (new)
2a. Emphasises that the geographical scope of such a framework should cover all the components of artificial intelligence, robotics and related technologies developed, deployed or used in the Union, including in cases where part of the technologies might be located outside the Union or not have a specific location;
Amendment 15 #
Draft opinion
Paragraph 2 a (new)
Paragraph 2 a (new)
2a. Underlines the need of constant monitoring of the use of AI; especially from the point of view of its advantages and disadvantages, as well as, its impact on the protection of Universal Human Rights;
Amendment 16 #
Draft opinion
Paragraph 2 b (new)
Paragraph 2 b (new)
2b. Recalls that the impact of AI systems should be considered not only from an individual perspective but also from the perspective of a society as a whole; calls to fully incorporate in a new framework, the human-centric approach based on the Communication on Building Trust in Human-Centric AI and the input obtained in the Ethics Guidelines prepared by the High-Level Expert Group on AI;
Amendment 24 #
Draft opinion
Paragraph 3 a (new)
Paragraph 3 a (new)
3a. Calls for establishing of synergies and networks between the various European research centres on AI as well as other multilateral fora, such as: Council of Europe, the United Nations Educational Scientific and Cultural Organization (UNESCO), the Organisation for Economic Co-operation and Development’s (OECD),the World Trade Organisation and the International Telecommunications Union (ITU), in order to align their efforts and to better coordinate the developments of the AI technology;
Amendment 36 #
Draft opinion
Paragraph 4
Paragraph 4
4. Highlights that, based on a human- centric approach, the Unionwhereby technology fully respects Human Rights and humans retain authority over automated decision- making systems, the Union needs a robust AI regulatory framework focused on security and defence, followsing a path of responsibility, of protecting our citizens, and of defending our values, that its policies aim at preserving peace, preventing conflicts and strengthening international security, whilst seizing the opportunities that those technologies offer;
Amendment 43 #
Draft opinion
Paragraph 4 a (new)
Paragraph 4 a (new)
4a. Calls on the Member States and the European Commission to ensure that the algorithms used in defence systems, while keeping the necessary confidentiality, are governed by the principle of transparency, including a clear liability regime for the results of AI use; underlines that these algorithms must be constantly adjusted to the progress in AI technologies;
Amendment 47 #
Draft opinion
Paragraph 5
Paragraph 5
5. Underlines that the Union must be at the forefront of mastering those technologies by establishing well defined processes for their use, for understanding the related ethical aspects and for fostering an effective international regulatory framework that contains the inherent risks of these technologies and prevents use for malicious purposes; the Union working together with the Member States must determine the appropriate liability regimes applicable to innovations in AI and other immersive technologies in the field of security and defence thus establishing a legal basis for accountability and traceability mechanisms, those include in particular unintended harm to persons, be it material or immaterial, such as breach of fundamental rights;
Amendment 57 #
Draft opinion
Paragraph 6
Paragraph 6
6. Stresses that for any defence application of AI enabled systems, the Union should set technical and organisational standards to ensure their resilience against cyber-attacks and digital influence, as well as their compliance with the highest possible trustworthiness standards as regards the collection and exploitation of operational datareliability standards, active monitoring and supervision as regards the collection and exploitation of operational data; AI systems and applications intended to extract and synthesise data, and extrapolate results therefrom to inform decisions for matters relating to defence and national security, must be specific in scope and comply with the provisions set out in the current regulatory framework in terms of gathering and processing data; stresses that AI applications designed to process data for intelligence purposes in defence related activities should comply with data processing standards to avoid risks of unintended surveillance or infringement of individual rights; believes that for high-risk applications of AI-enabled technologies like facial recognition which lack a definitive regulatory framework at the EU level, the Union must ensure that their development and deployment is rightful, proportional and respects the rights of individuals; stresses the importance of monitoring competent national law enforcement authorities which develop and deploy AI-enabled systems and technologies to maintain public order so as to mitigate the disproportional risks of predictive policing;
Amendment 59 #
Draft opinion
Paragraph 6
Paragraph 6
6. Stresses that for any defence application of AI enabled systems, the Union should set technical and organisational standards to ensure their resilience against cyber-attacks and digital influence, as well as their compliance with the highest possible trustworthiness standards as regards the collection and exploitation of operational data; draws attention to the need of careful analysis of the algorithms on which AI makes its decisions; emphasises the importance for transparency and accountability of AI algorithms; notes the important distinction between transparency of algorithms and transparency of the use of algorithms;
Amendment 65 #
Draft opinion
Paragraph 6 a (new)
Paragraph 6 a (new)
6a. Notices the great risk of AI in the area of disinformation; underlines that, if not regulated, AI technologies might have ethically adverse effects by exploiting bias in data and algorithms that may lead to disinformation, creating information bubbles and exploiting biases incorporated into AI algorithms;
Amendment 66 #
Draft opinion
Paragraph 6 a (new)
Paragraph 6 a (new)
6a. Calls on the Commission and the Member States to work together in establishing a framework to prohibit the activation of AI-enabled systems in the field of security and defence using advanced data to carry out lethal or destructive actions without a human-in- the-loop-principle;
Amendment 68 #
Draft opinion
Paragraph 7
Paragraph 7
7. Highlights the need to adopt clear safety and security provisions and requirements, with proper certifications, for AI-systems in defence, and carry ourt constant monitoring, regular tests and verifications across the entire life cycle; underlines the necessity of ensuring compliance with applicable standards and obtained certifications where AI modifies e.g. through machine learning the functionality and behaviour of systems in which it is integrated, in order to ensure full traceability of decisions made with involvement of AI;
Amendment 70 #
Draft opinion
Paragraph 7 a (new)
Paragraph 7 a (new)
7a. Calls on the Commission to embed cybersecurity capacity-building in its industrial policy in order to ensure the development and deployment of safe, resilient and robust AI-enabled and robotic systems; calls on the Commission to explore the use of blockchain-based cybersecurity protocols and applications to improve the resilience, trust and robustness of AI infrastructures through disintermediated models of data encryption; encourages European stakeholders to research and engineer advanced features that would facilitate the detection of corrupt and malicious AI- enabled &robotics systems which could undermine the security of the Union and of citizens;
Amendment 72 #
Draft opinion
Paragraph 7 a (new)
Paragraph 7 a (new)
7a. Highlights the fact that AI framework in defence and security should develop benchmarks for ethically responsible and accepted uses of AI technologies; underlines that these criteria must be constantly adjusted to the progress in AI technologies;
Amendment 74 #
Draft opinion
Paragraph 7 b (new)
Paragraph 7 b (new)
7b. Emphasises the need for an embedded option to reduce an AI-enabled system’s computations to a form comprehensible by humans throughout a given system’s lifecycle; considers that AI-enabled systems, products and technology purposed for military use should be equipped with a ‘black box’ to record every data transaction carried out by the machine, including the logic that contributed to its decisions, as well as with a 'switch-off' button which would instantly enable humans to deactivate the AI-enabled system;
Amendment 81 #
Draft opinion
Paragraph 8
Paragraph 8
8. Stresseds that all AI-systems in defence must have a concrete and well- defined domain of use and must be endowed withwhereby humans retain the abilitgency to detect and, disengage or diseactivate deployed systems should they move from their domain of use or engage in any escalatory or unintended action;
Amendment 85 #
Draft opinion
Paragraph 9
Paragraph 9
9. Underlines that the entire responsibility for the decision to design, develop, deploy and use AI-systems must rest on human operators and, as there must be meaningful human control over any weapons system and human intent in the decision to use force; underlines that the human-in- the-loop principle must also be applied to the command and control of AI- enabled systems; stresses that AI-enabled systems must allow the military leadership to assume its full responsibility and exercise the necessary level of judgmentaccountability for the use of lethal force and exercise the necessary level of judgment, which cannot be endowed to machines as it must be based on distinction, proportionality and precaution, for taking lethal or large-scale destructive action bey means of such systems; recalls in this respect its position on a ban on the development, production and use of fully autonomous weapons systems enabling strikes to be carried out without meaningful human intervention;
Amendment 94 #
Draft opinion
Paragraph 9 a (new)
Paragraph 9 a (new)
9a. Stresses that the EU must take the lead in promoting the establishment of international norms regarding the ethical and legal parameters of the development and use of fully autonomous, semi- autonomous and remotely operated lethal weapons systems; Member States should develop national strategies for the definition, status and use of lethal autonomous weapons (LAWs) towards a comprehensive strategy on the EU level;
Amendment 98 #
Draft opinion
Paragraph 9 b (new)
Paragraph 9 b (new)
9b. Calls on the Commission to work together with Member States’ national competent authorities and other stakeholders participating in the development and deployment of AI- enabled systems, products and technologies to establish a safe, secure and resilient framework whereby the source code of AI-enabled systems is shared, monitored and verified to mitigate potential deviations from the governing principles and ethical framework underpinning AI technology in the field of security and defence; suggests to the Commission that the EU must retain ownership of the intellectual property of EU-funded research on AI-enabled systems, products and technologies in security and defence;
Amendment 103 #
Draft opinion
Paragraph 10
Paragraph 10
10. Underlines that the Union must promote understanding of the military implications of AI, of robotics and of autonomy; considers that the Union needs to promote the acquisition of the necessary skills and knowledge on technology development processes and operational methods throughout the supply chain and over the full lifecycle of AI-enabled military capabilities;
Amendment 105 #
Draft opinion
Paragraph 10 a (new)
Paragraph 10 a (new)
10a. Believes that enhanced cooperation between Member States and the Commission is necessary to guarantee coherent cross-border rules in the Union to encourage the collaboration between European industries and allow the development and deployment of AI- enabled technologies consistent with the prescribed safety and security standards, and the ethical framework governing the development and deployment of AI technology;
Amendment 108 #
Draft opinion
Paragraph 10 a (new)
Paragraph 10 a (new)
10a. Underlines, that despite its added value, AI comes with a number of weaknesses, one of them relating to all sorts of different types of biases; emphasises that AI technologies should clearly be void of any sort of profiling, especially regarding gender;
Amendment 116 #
Draft opinion
Paragraph 11
Paragraph 11
11. Stresses the need to overcome the current fragmentation within the Union as regards national AI-related law, research, innovation and expertise in the area of AI, which puts in jeopardyendangers the functioning of the internal market and the objective to ensure trustworthyreliable and secure development of AI in Europe; in this respect welcomes the inclusion of AI- related projects under the European Industrial Development Programme (EDIDP); believes that the future European Defence Fund (EDF) and the Permanent structured cooperation (PESCO) also offer well adapted frameworks for future AI- related projects that would help to better streamline Union efforts in this field; stresses that AI-related projects should be synchronized with the wider EU civilian programmes devoted to AI; notes that in line with the European Commission’s White Paper on AI (COM2020/65final) excellence and testing centres concentrating on research and development of AI in the field of security and defence should be established with vigorous specifications underpinning the participation of and investment from private stakeholders;
Amendment 121 #
Draft opinion
Paragraph 11 a (new)
Paragraph 11 a (new)
11a. Emphasises that the development of AI that respects fundamental rights and supports the public interest requires the strategic pooling and sharing of data in the EU between private and public entities, as well as the strengthening of an EU AI ecosystem, which involves public, private, and civil society stakeholders; calls on the European Commission to foster dialogue among Member States, researchers, academics, civil society actors and the private sector so as to have inclusive policymaking processes when it comes to defence-related AI regulations;
Amendment 126 #
Draft opinion
Paragraph 11 a (new)
Paragraph 11 a (new)
11a. Calls on the Commission to present their "Reinforcement of the Skills Agenda", announced in the White Paper on Artificial Intelligence on the 19th February2020, as soon as possible - in order to ensure that everyone in Europe can benefit from the digital transformations of the EU economy;
Amendment 129 #
Draft opinion
Paragraph 11 b (new)
Paragraph 11 b (new)
11b. Stresses the importance of the creation of an ethical code of conduct underpinning the deployment of weaponised AI-enabled systems in military operations, similar to the existing regulatory framework prohibiting the deployment of chemical and biological weapons; is of the opinion that the Commission should initiate the creation of standards on the use of AI-enabled weapons systems in warfare according to international humanitarian law;
Amendment 133 #
Draft opinion
Paragraph 12
Paragraph 12
12. Calls on the European Commission and on the VP/HR to present, also as part of an overall approach, a sectoral AI strategy for defence-related activities within the Union framework, that ensures both respect for citizens’ rights and EU’s strategic interests that should propose a consistent regulatory approach spanning from the inception of AI-enabled systems to their military uses; calls on these regulatory efforts to be supported by meaningful monitoring schemes, so that normative frameworks are not outplaced by technological development and new methods of warfare; calls on the Council, the European Commission and on the VP/HR to enter in a structured dialogue with the European Parliament to that end.