Weekly AI Governance Brief: 8–14 June 2026
Bringing you the latest developments in the AI governance world.
European Commission publishes final code of practice for AI-generated content transparency
On 10 June 2026, the European Commission and the AI Office published the final Code of Practice on Transparency of AI-Generated Content. The code is intended to support compliance with Article 50 of the EU AI Act, which concerns transparency obligations for providers and deployers of certain AI systems, including generative AI systems.
The Commission states that the underlying Article 50 obligations become applicable from 2 August 2026. The code itself is voluntary, but the legal duties it supports are not. It addresses obligations relating to the marking and labelling of AI-generated content, detectability, deepfake labelling, and certain AI-generated text published on matters of public interest.
The code is structured in two sections, one aimed at providers and one at deployers. The Commission has also published EU icons for labelling AI-generated content. According to the Commission, the code is undergoing an adequacy assessment by the Commission and the AI Board. If the code is positively assessed, signatories will be able to rely on its measures to demonstrate compliance. Non-signatories will need to show that their alternative measures are adequate.
Why this matters
This is a practical implementation instrument ahead of the 2 August 2026 application date for the AI Act’s transparency obligations. It gives providers and deployers a clearer operational reference point for marking, labelling, and detecting AI-generated content.
The governance relevance is especially direct for organisations deploying generative AI tools in public communication, platform services, media production, customer interaction, or public-interest content. The code does not remove the need for internal assessment, but it creates a more concrete basis for compliance design. It also clarifies the institutional role of the Commission and AI Board in assessing whether voluntary measures are adequate for demonstrating compliance with binding legal obligations.
Germany’s Bundestag adopts national AI Act implementation bill
On 11 June 2026, the German Bundestag adopted the federal bill to implement the EU AI Act. The Bundestag states that it considered the government bill, Bundestag Drucksache 21/4594, and adopted it in the committee-amended version, Drucksache 21/6407. The Federal Government’s account, published on 12 June 2026, presents the law as Germany’s vehicle for designating national authorities required under the EU AI Act and for creating a national enforcement structure.
According to the Federal Government, the bill specifies which federal and Länder authorities will be responsible for market surveillance and notification of AI systems. It also regulates cooperation among authorities and creates clearer contact points for companies.
A central element is the role assigned to the Bundesnetzagentur. The Government states that a coordination and competence centre is to be established there. It also says that the Bundesnetzagentur is to set up at least one AI regulatory sandbox. The Government notes that the bill still requires approval by the Bundesrat.
Why this matters
This is a significant national implementation step because it connects EU-level AI Act obligations to domestic supervisory architecture. For organisations operating in Germany, the bill provides a clearer view of which authorities may be involved in market surveillance, notification, coordination, and sandbox-based testing.
The practical governance relevance lies in institutional responsibility. AI Act compliance will not be managed only through EU-level rules. It will also depend on national authorities, cooperation mechanisms, and administrative contact points. Germany’s implementation bill, therefore, helps define the route through which companies may interact with supervisory bodies in one of the EU’s largest markets.
EU and Brazil formalise digital-governance cooperation, including AI and platform-risk issues
On 12 June 2026, the European Commission announced that the EU and Brazil would sign a Digital Partnership. The Commission describes the partnership as covering data governance, artificial intelligence, digital infrastructure and connectivity, online platforms, and digital public goods and services.
The Commission presents Digital Partnerships as a central instrument of the EU’s external digital policy. In this case, it states that the partnership is intended to contribute to a more inclusive and rules-based global digital governance system.
In the same context, Commission services responsible for enforcing the Digital Services Act signed an administrative arrangement with Brazil’s data protection authority, ANPD. The arrangement focuses on protecting minors online. The Commission says it covers transparency obligations for digital platforms, risk assessment and mitigation measures, and technological cooperation on algorithms and artificial intelligence. It also provides for technical exchanges, joint studies, and research cooperation.
Why this matters
The immediate relevance is not a new domestic compliance obligation within the EU. Its importance lies in regulatory cooperation and institutional alignment beyond the EU’s borders. The development shows how AI, platform governance, online safety, and data protection are increasingly being addressed through linked cooperation mechanisms.
For organisations operating across jurisdictions, the practical relevance is that digital governance issues are being framed through overlapping institutional channels. Platform-risk management, algorithmic transparency, and AI-related cooperation are not being treated as isolated policy areas. They are being incorporated into broader digital-governance partnerships and regulator-to-regulator arrangements.
Canada introduces Bill C-34, extending online-safety duties to chatbot services
On 10 June 2026, Canada introduced Bill C-34 in the House of Commons, and the bill received first reading. The bill would enact the Digital Safety Act and the Digital Safety Commission of Canada Act.
The bill summary states that it would apply to operators of regulated social media services, regulated chatbot services, and other regulated online services. Its stated purpose is to promote safety, particularly for children, reduce harm from harmful online content, and ensure transparency and accountability.
The bill would impose duties on regulated chatbot services. These include measures to mitigate the risk that a chatbot communicates harmful content, intervention in crises, and measures to mitigate the risk that a chatbot engages in certain harmful behaviours. The bill also contemplates broader duties for regulated services, including child-protection design features, age-related controls, transparency obligations, and the filing of a digital safety plan with the proposed Digital Safety Commission of Canada.
The Government of Canada’s announcement states that the proposed legislation would create safety requirements for AI chatbot services and establish the new Commission as the enforcement body.
Why this matters
This development is relevant because it places chatbot services directly within online-safety legislation. It shows a legislative approach in which AI chatbot governance is connected to harmful-content controls, child protection, crisis intervention, and transparency duties.
For compliance teams, the practical point is the combination of service design, risk mitigation, reporting, and institutional supervision. The proposed Digital Safety Commission would create a dedicated enforcement body, while regulated services would be expected to manage chatbot-related harms through defined operational duties. Although this is a Canadian bill, it is relevant for organisations tracking how online-safety frameworks are being extended to AI-enabled services.
UK MHRA launches real-world regulatory sandbox for AI medical devices
On 10 June 2026, the UK Medicines and Healthcare products Regulatory Agency announced the launch of London Region I, a regulatory sandbox for AI-enabled medical devices. The MHRA describes the sandbox as a secure, real-world environment developed with NHS England, London, and the London Health Innovation Networks.
The sandbox is intended to support safe deployment of AI-enabled medical devices while generating additional evidence of benefit across NHS settings. According to the MHRA, up to 10 AI medical device manufacturers will be selected for the initial phase. They will work with NHS providers in live clinical settings under MHRA oversight.
The agency states that the sandbox is intended to generate robust real-world evidence on safety and effectiveness while creating a clearer route to wider adoption. Expressions of interest from NHS providers and device manufacturers are due to open the following month.
Why this matters
This is a sector-specific governance development with direct relevance to medical-device regulation and AI deployment in clinical settings. It frames AI oversight as an evidence-generation process in supervised real-world environments, rather than only as a static pre-market assessment.
For organisations developing or deploying AI-enabled medical devices, the operational significance lies in the connection between regulatory oversight, clinical deployment, and post-market evidence. The sandbox structure also clarifies the role of the MHRA in supervising the generation of safety and effectiveness evidence within live healthcare settings.
New York disclosure law for AI-generated synthetic performers in advertising takes effect
On 9 June 2026, Governor Kathy Hochul announced that New York’s law requiring disclosure when advertisements include AI-generated synthetic performers was in effect. The Governor’s office states that the law requires persons who produce or create an advertisement to identify when the advertisement includes such a synthetic performer.
The underlying New York State Senate bill page for 2025-S8420A states that the law requires advertisements to disclose the use of a synthetic performer. It also provides for a civil penalty of $1,000 for a first violation and $5,000 for subsequent violations. The law is reflected in New York General Business Law § 396-b.
Why this matters
The law is narrow in scope, but it creates a live transparency obligation for synthetic media in advertising. Its compliance relevance lies in production and disclosure workflows for commercial content that uses AI-generated performers.
For organisations producing advertisements, commissioning campaigns, or using synthetic likenesses, the operational issue is whether disclosure requirements are built into content creation and approval processes. The law also illustrates how synthetic-content governance can be addressed through targeted commercial transparency rules, rather than only through broad AI legislation.
Looking ahead
The developments this week show AI governance moving through implementation instruments, national supervisory design, and sector-specific oversight mechanisms. The EU items focus on practical compliance pathways under the AI Act, national authority designation, and external digital-governance cooperation. The international items show AI-related obligations being integrated into online safety, healthcare regulation, and advertising transparency.
Across the week, the common feature is not the creation of broad new AI principles. It is the translation of governance expectations into institutional roles, compliance tools, and operational duties. This is visible in the Commission’s transparency code, Germany’s implementation bill, Canada’s chatbot provisions, the MHRA sandbox, and New York’s advertising disclosure rule.
Sources
European Commission, Code of Practice on Transparency of AI-Generated Content: https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content
European Commission, Commission publishes Code of Practice on marking and labelling AI-generated content: https://digital-strategy.ec.europa.eu/en/news/commission-publishes-code-practice-marking-and-labelling-ai-generated-content
Deutscher Bundestag, Ja zur Durchführung der Verordnung über künstliche Intelligenz: https://www.bundestag.de/dokumente/textarchiv/2026/kw24-de-ki-verordnung-1113452
Federal Government of Germany, KI-Verordnung beschlossen: https://www.bundesregierung.de/breg-de/aktuelles/ki-verordnung-beschlossen-2360482
European Commission, EU and Brazil deepen ties through Digital Partnership: https://digital-strategy.ec.europa.eu/en/news/eu-and-brazil-deepen-ties-through-digital-partnership
European Commission services, Commission services cooperate with Brazil’s Data Protection Agency on protection of minors online: https://digital-strategy.ec.europa.eu/en/news/dsa-br
Parliament of Canada, Government Bill C-34, Safe Social Media Act, first reading: https://www.parl.ca/DocumentViewer/en/45-1/bill/C-34/first-reading
UK Medicines and Healthcare products Regulatory Agency, Pioneering AI health innovations regulatory sandbox launched: https://www.gov.uk/government/news/pioneering-ai-health-innovations-regulatory-sandbox-launched
Governor of New York, Governor Hochul Announces First-in-the-nation Law Requiring Disclosure When Advertisements Include AI-generated Synthetic Performers is in Effect: https://www.governor.ny.gov/news/governor-hochul-announces-first-nation-law-requiring-disclosure-when-advertisements-include-ai
New York State Senate, Senate Bill 2025-S8420A: https://www.nysenate.gov/legislation/bills/2025/S8420/amendment/A