The EU Artificial Intelligence Act (AIA) establishes different legal principles for different types of AI systems. While prior work has sought to clarify some of these principles, little attention has been paid to robustness and cybersecurity. This paper aims to fill this gap. We identify legal challenges and shortcomings in provisions related to robustness and cybersecurity for high-risk AI systems (Art. 15 AIA) and general-purpose AI models (Art. 55 AIA). We show that robustness and cybersecurity demand resilience against performance disruptions. Furthermore, we assess potential challenges in implementing these provisions in light of recent advancements in the machine learning (ML) literature. Our analysis informs efforts to develop harmonized standards, guidelines by the European Commission, as well as benchmarks and measurement methodologies under Art. 15(2) AIA. With this, we seek to bridge the gap between legal terminology and ML research, fostering a better alignment between research and implementation efforts.

While such provisions in the directive could make businesses "unhappy", Sarah Cameron, technology legal director at law firm Pinsent Masons, said the rules helped clarify liability for AI-enabled products for consumers and businesses alike.

For some time, there has been an acute need for a legal framework to govern artificial intelligence (AI). This is largely due to the number of longstanding regulatory and ethical concerns surrounding the technology since its inception. I am a firm believer that we need to properly govern AI to prevent issues such as unethical biases, the undermining of legal and regulatory norms, and the blurred lines of organizational accountability from happening. These problems can seriously overwhelm users, business and citizens, and yet would be so avoidable if proper governance for AI was in place. So, earlier this year, when the EU Commission put forward the idea of a world-first legal framework for AI, great progress was made.

In March this year, the UK government announced an assertive agenda on artificial intelligence (AI) by launching a UK Cyber Security Council and revealing plans to publish a National Artificial Intelligence Strategy (the UK Strategy). The details of the UK Strategy will be released later this year, but at this point we understand that it will focus in particular on promoting growth of the economy through widespread use of AI with, at the same time, an emphasis on ethical, safe, and trustworthy development of AI--including through the development of a legislative framework for AI which will promote public trust and a level playing field. Shortly after the UK government's announcement, the EU Commission published a proposed EU-wide AI legislative framework (the EU Regulation) which is part of the Commission's overall "AI package". The EU Regulation is focused on ensuring the safety of individuals and the protection of fundamental human rights, and categorises AI into unacceptable, high- or low-risk use cases. The EU Regulation proposes to protect users "where the risks that the AI systems pose are particularly high". The definition and categories of high-risk use cases of AI are broad, and capture many if not most use cases that relate to individuals, including AI use in the context of biometric identification and categorisation of natural persons, management of critical infrastructure, and employment and worker management.

A written proposal to ban several uses of artificial intelligence (AI) and to place new oversight on other "high-risk" AI applications--published by the European Commission this week--met fierce opposition from several digital rights advocates in Europe. Portrayed as a missed opportunity by privacy experts, the EU Commission's proposal bans four broad applications of AI, but it includes several loopholes that could lead to abuse, and it fails to include a mechanism to add other AI applications to the ban list. It deems certain types of AI applications as "high-risk"--meaning their developers will need to abide by certain restrictions--but some of those same applications were specifically called out by many digital rights groups earlier this year as "incompatible with a democratic society." It creates new government authorities, but the responsibilities of those authorities may overlap with separate authorities devoted to overall data protection. Most upsetting to digital rights experts, it appears, is that the 107-page document (not including the necessary annexes) offers only glancing restrictions on biometric surveillance, like facial recognition software.

This article is brought to you in association with the European Parliament. MEPs approved proposals to address long-term opportunities and legal challenges posed by AI in the area of ethics, civil liability and intellectual property. The Legal Affairs Committee adopted three reports on Thursday on specific issues linked to the increased development and use of artificial intelligence systems. The Commission is expected to put forward a legislative proposal on the matter in early 2021. The legislative initiative by Iban García del Blanco (S&D, ES), adopted with 20 votes in favour, none against and 4 abstentions, urges the EU Commission to present a new legal framework outlining the ethical principles to be used when developing, deploying and using artificial intelligence, robotics and related technologies in the EU, including software, algorithms and data.

The Legal Affairs Committee adopted three reports on Thursday on specific issues linked to the increased development and use of artificial intelligence systems. The Commission is expected to put forward a legislative proposal on the matter in early 2021. The legislative initiative by Iban García del Blanco (S&D, ES), adopted with 20 votes in favour, none against and 4 abstentions, urges the EU Commission to present a new legal framework outlining the ethical principles to be used when developing, deploying and using artificial intelligence, robotics and related technologies in the EU, including software, algorithms and data. MEPs adopted proposals on several guiding principles that must be taken into account by future laws including a human-centric, human-made and human-controlled AI; safety, transparency and accountability; safeguards against bias and discrimination; right to redress; social and environmental responsibility, and respect for fundamental rights. When it comes to AI with machine-learning (self-improving) capacities, it should be designed to allow for human oversight.

Organising ethical debates has long been an efficient way for industry to delay and avoid hard regulation. Europe now needs strong, enforceable rights for its citizens, writes Green MEP Alexandra Geese. If the rules are too weak, there is a too great a risk that our rights and freedoms will be undermined: This currently applies to all applications of artificial intelligence, which up to now have only been based on non-binding ethical principles and values. In this legislation, Europe has the chance to adopt a legal framework for AI with clear rules. We need strong instruments to protect our fundamental rights and democracy.

In the last two months, i.e. in April and May 2019, both the EU Commission and the OECD published guidelines for trustworthy and ethical Artificial Intelligence (AI). In both cases, these are only guidelines and, as such, are not legally binding. Both sets of guidelines were compiled by experts in the field. Let's have a closer look. "Why do we need guidelines for trustworthy, ethical AI?" you may ask.

In the last two months, i.e. in April and May 2019, both the EU Commission and the OECD published guidelines for trustworthy and ethical Artificial Intelligence (AI). In both cases, these are only guidelines and, as such, are not legally binding. Both sets of guidelines were compiled by experts in the field. Let's have a closer look. "Why do we need guidelines for trustworthy, ethical AI?" you may ask.