AI Strategies in Czech Republic

Yes. The Government of the Czech Republic has commissioned academic institutions to prepare an expert report mapping the potential of AI development in the Czech Republic. 

Furthermore, an expert group called “AI Observatory and Forum” (AIO&F), a joint project by the Government and the Ministry of Industry and Trade, will identify legal obstacles during the development, research and use of artificial intelligence and formulate proposals on how to eliminate them.

N/A

No draft regulations have been prepared to date. However, the most relevant papers include:

• “Memorandum of cooperation on the development of AI”, signed by the Government of the Czech Republic and industry representatives on 16 January 2019. The main goal is cooperation between the parties on preparing a comprehensive national strategy for AI;
• “Analysis of the Development Potential of Artificial Intelligence (the “Analysis of AI”), which evaluates both current and future potential of the usage of AI technologies in the Czech Republic, including an analysis of its socio-economic impacts and the influence of the legal and ethical environment in this process. It also anticipates the preparation of a national strategy for AI;
• Innovation Strategy of the Czech Republic 2019–2030 introduced on 14 February 2019 sets out new priorities, which, if achieved, should put the Czech Republic among the most innovative countries in Europe by 2030. It explicitly states that the aim of the Czech Republic is to prepare society for trends such as AI and presents tools for achieving the goal (e.g. financial support for innovative AI solutions via RD&I programmes, launch of a long-term communications campaign using examples of scientific and commercial success, the success of scientific teams and their leaders in the field of AI).
• National Artificial Intelligence Strategy (NAIS) adopted on 7 May 2019 by the Czech Government. The main aim of this strategy is for the Czech Republic to become a “model country for AI” by the time of its Presidency of the Council of the EU in 2022;
• Regulatory Framework for Artificial Intelligence in the European Union, published on 26 November 2019. This document emphasises that “securing the safety of citizens is the key precondition for the true implementation of fundamental human rights and freedoms”;
• “Visegrad 4 countries” paper (“V4 Document”), which contains opinions on AI and maximising its benefits ahead of the release of the European Commission’s communication on the topic on the Digital day on 10 April 2018.

The definition of AI has not yet been set down; however, it is likely that the definition  the HLEG proposed in the document Definition of Artificial Intelligence will be used as a basis.

The Czech Republic does not have any clear stance on the concept of AI having a legal personality. AI is currently considered as software, i.e. a literary work from the copyright law perspective. Thus, granting a legal personality to robots is not currently possible under Czech law.

The liability for damage caused by AI cannot currently be subsumed under any specific provision of Czech law; however, the applicability of provisions on strict liability for damage is discussed, whether the liability is governed by liability for damage caused by a defect in an item (product) or damage resulting from operating activities, and other provisions.

The Czech Republic has not presented any detailed concepts of how the liability of AI should be regulated. The Analysis of AI states that the Czech Republic may find inspiration in “Liability for Artificial Intelligence and Other Emerging Digital Technologies”, a paper by the Expert Group on Liability and New Technologies published in 2019. The paper lays down several options of approach to liability, i.e. liability of operator of the technology (e.g. driver of an autonomous vehicle) and liability of the producer of the technology. 

So far there are no detailed concepts on the safety compliance of AI systems.

The Analysis of AI suggests adopting certification of AI or engaging in all activities related to certification at the EU level.

The Analysis of AI suggests adopting the GDPR interpretation rules to clarify the scope of the obligation to provide information related to the process used during automated decision-making, including profiling. The interpretation should respect not only the rights of data subjects but also rights to the protection of the intellectual property of data controllers.

The analysis further emphasises the principle of privacy-by-design, use of data protection impact assessment and the use of methods reducing the amount of data necessary for the AI to learn or methods using encryption to ensure the protection of privacy.

The Analysis of AI suggests adopting technical means of distinguishing between personal and non-personal data, and refers to the EU approach.

There is no specific guideline regarding the IP created by AI.

However, as AI does not have a legal personality, it cannot be considered as the author or inventor of IP, and therefore AI can only serve as a tool which is used to create IP. 

The Analysis of AI suggests amending the Copyright Act to ensure the protection of works created by AI.

The Analysis of AI recommends the Czech Republic join the effort at the EU level and take part in forming ethical rules for trustworthy AI.

The Analysis of AI emphasises human oversight. Supervision may be key to identifying legal liability of the AI or it may serve as an oversight of its functioning.

The Analysis of AI does not propose any specific solutions; however, it does acknowledge that AI could have impacts on competition and highlights potential problems, e.g. algorithmic collusion and price discrimination, based on consumer behaviour.

The Analysis of AI recommends setting down the obligation to inform an individual person that he/she is communicating with AI, e.g. a chatbot.

The Analysis of AI states that contracts related to selling AI or providing AI services may have adverse consequences for consumers due to the misuse of adhesion contracts (i.e. form contracts). The Analysis of AI suggests that preparing contract templates might solve this challenge.

The Analysis of AI considers the impacts of AI in four main areas related to employment: (a) social security; (b) qualification increase for certain professions; (c) education and training of a new workforce; and (d) development of a wider system of public policies related to AI. 

The Analysis of AI refers to the existing regulation of algorithmic trading on capital markets and the growing use of “robo-advisors”, i.e. applications advising clients on trading in the capital markets, and foresees a great future for AI in this area. However, it notes that the concept of liability for damage and public regulation have to be adjusted to fully reflect the challenges.

The Analysis of AI notes that AI is to some extent already used by social security authorities to calculate benefits. The analysis emphasises that human control should still be present (at least during the revision of a decision). However, no solutions were proposed.

One of the NAIS medium-term objective (implemented by 2027) is the introduction of AI elements in the judiciary, such as the sounding of all recording halls, the use of spoken word transcripts in selected agendas, and the involvement of artificial intelligence elements in the justice anonymizer. 

One of the NAIS long-term objectives (implemented by 2035) is the e.g extension of AI elements in the judiciary, such as use of transcription in machine form in all court cases, deployment of an anonymization tool using artificial intelligence elements, automatic indexing of court decisions allowing easier search among decisions or involvement of artificial intelligence for handling requests of users and citizens.