AI Law

AI Act in practice: workshop recap

AI Act in practice: workshop recap

AI Act obligations depend on how a company uses AI and its role. At our workshop at Kampus Hybernská, we explored how to identify those obligations and address them in contracts, internal policies and day-to-day operations.

On Thursday 12 November 2026, we held the workshop “AI Act in practice: What innovative companies need to know” at Kampus Hybernská in Prague. We organised it with prg.ai and B4I (brain4industry) as part of the fifth Dny AI 2026 festival. The event was co-funded by the EU, with the Czech Ministry of Industry and Trade as a partner. Jaroslav Menčík, Partner, and Bohdan Zubač, Associate, opened the expert programme.

What we covered

Our opening presentation focused on the issues companies need to address today.

Jaroslav Menčík speaking at the AI Act in practice workshop
Jaroslav Menčík, Partner
  • Risk and the company’s role. The AI Act distinguishes between prohibited practices, high-risk systems, systems with transparency obligations and general-purpose AI models. Obligations depend on whether a company acts as a provider, deployer, importer or distributor. The regulation also applies to entities outside the EU where their system’s output is used in the EU.
  • High-risk systems. Examples include AI used in recruitment, creditworthiness assessment or as a safety component of regulated products such as medical devices. Providers’ obligations cover risk management, documentation and conformity assessment. Deployers must use the system in line with its instructions, ensure human oversight and input data quality, monitor its operation and keep logs. Banks and insurers must assess the impact on fundamental rights before deployment.
  • Transparency. A chatbot must tell users they are interacting with AI unless this is obvious. Providers must mark synthetic content in a machine-readable format, and deployers must disclose that deepfakes are artificially generated. Not every AI-generated text needs a label: the duty covers texts on matters of public interest that have not undergone human editorial review. The information must be clear, provided no later than the first interaction and not buried in terms and conditions.
  • Deadlines. Prohibitions and the AI literacy duty have applied since 2 February 2025, and the rules for general-purpose AI models since 2 August 2025. The next deadline is 2 December 2026: the so-called AI Digital Omnibus introduces two new prohibitions from that date (non-consensual sexual deepfakes of an identifiable person and material depicting child sexual abuse). The transition period also ends for machine-readable marking of outputs from generative AI placed on the market before 2 August 2026. Obligations for high-risk systems follow on 2 December 2027 (Annex III) and 2 August 2028 (Annex I).
  • The GDPR and other laws. The AI Act does not replace the GDPR. Where an AI system processes personal data, both regimes apply. We recommend a single process for managing data and technology risks rather than separate “AI” and “GDPR” audits. The area of use often matters more than the technology itself.
  • Czech implementation. The Ministry of Industry and Trade is preparing an Artificial Intelligence Act and an amendment to the Market Surveillance Act; neither has been promulgated yet. Under the draft, the Czech Telecommunication Office would supervise the market, and the regulatory sandbox would be run by the Office for Standards, Metrology and Testing (ÚNMZ) and the Czech Standardisation Agency.

Perspectives from practice

Guest speakers followed the legal session. Ondřej Kramoliš, Senior Legal Counsel at Gen Digital, covered AI governance in a global technology company and the legal and operational questions it raises. Monika Nguyen, Account & Project Manager at Cocoon, focused on a design and branding agency’s experience with generative AI, including the use of references and transparency around AI-generated content. Petr Voplakal of the Central Bohemian Innovation Centre covered support for AI implementation through the European EDIH programmes, including examples of proof-of-concept and proof-of-technology projects.

Ondřej Kramoliš of Gen Digital speaking at the AI Act in practice workshop
Ondřej Kramoliš, Gen Digital

The morning closed with questions and discussion, followed by networking.

Six steps for companies

We ended our presentation with six practical steps any company can take:

  1. Take an inventory of AI tools. Include tools individual teams use informally.
  2. Classify the risks. Assess each use case separately: is it a prohibited practice, a high-risk system, or do transparency obligations apply?
  3. Determine the company’s role. Provider, deployer, importer, distributor or manufacturer of a product with integrated AI.
  4. Review contracts with AI suppliers. Check documentation, security, data, audit rights, incidents and liability.
  5. Set internal rules and provide training. Cover AI policy, tool approval, prohibited inputs, output review and content labelling.
  6. Document compliance and monitor changes. Compliance does not end when a system is deployed; models and the legal framework continue to evolve.
Petr Voplakal, Jaroslav Menčík and Bohdan Zubač on stage at Kampus Hybernská
Petr Voplakal (Central Bohemian Innovation Centre), Jaroslav Menčík and Bohdan Zubač

Next steps

To identify where your company uses AI and which obligations apply, see our AI law services, from an AI audit to AI policies and supplier contracts. For questions, contact Jaroslav Menčík or Bohdan Zubač. To hear about future events, sign up for our newsletter.

Ambit banners with a speaker's silhouette
AI ActArtificial intelligenceEvents
Author
Jaroslav Menčík

Jaroslav Menčík

Jaroslav Menčík is Partner and Head of the IT & Technology Team at Ambit, leading the firm's legal practice in information technology, software, and innovation-driven industries. With more than 15 years of experience, he plays a pivotal role in establishing Ambit as the foremost technology-focused law firm serving entrepreneurs, investors, and established companies across the CEE region.

Education
  • Charles University in Prague, Faculty of Law
    Doctor of Jurisprudence (JUDr.) · 2024
  • Charles University in Prague, Faculty of Law
    Doctor of Laws (Ph.D.) · 2023
  • New York University School of Law
    LL.M. (Hauser Global Scholar) · 2017
  • Czech University of Life Sciences
    Master of Economics and Management (Ing.) · 2015
  • Czech Bar Association
    Attorney (admitted with honors) · 2013
  • Masaryk University, Faculty of Law
    Master of Laws (Mgr.) · 2009
  • Aberystwyth University, UK
    International Program in Law · 2008
Ask directly

Have a question about this article?

Our team will get back to you within two working days.