EU’s AI Labeling Mandate Kicks In Next Year — But Text Gets a Pass
Published on 07/21/2026 at 23:41 | Redaktion boerse-global.de
Businesses that generate images, videos, or audio with artificial intelligence will be required to mark those outputs with machine-readable labels — such as watermarks or signed metadata — starting August 2, 2026, under detailed guidelines the European Commission released just ahead of the AI Act’s enforcement timeline.
Text-based content is exempt from the labeling requirement for now. The Commission says the technology to reliably detect AI-written text is not mature enough. Content that undergoes human editorial review also escapes the rule. Companies running existing AI systems have until December 2, 2026, to retrofit them for compliance.
Penalties for violations are steep: fines can reach €15 million or 3 percent of a company’s global annual revenue, whichever is higher.
Germany Picks Its AI Watchdog
Germany’s Bundestag passed the national legal framework on June 11, 2026, designating the Federal Network Agency (Bundesnetzagentur) as the central market surveillance authority. The agency will set up its own coordination and competence center for AI oversight.
The legislation introduces AI real-world testing labs and a registry for high-risk applications used in critical infrastructure. The Bundesrat, Germany’s upper house, still needs to approve the law.
Legal experts caution that the AI Act does not replace the General Data Protection Regulation (GDPR). Companies must continue documenting how AI systems process personal data and, for high-risk uses, conduct a data protection impact assessment. Since February 2025, Article 4 of the AI Act has also required general AI literacy among staff.
Keeping up with overlapping regulations like the AI Act and GDPR demands rigorous documentation. Your workplace safety processes deserve the same level of care. A free Health & Safety Toolkit provides ready-to-use risk assessments and checklists aligned with UK law, helping you protect your team and stay compliant. Download the free Health & Safety Toolkit
86% of Firms Pilot AI Agents — But Trust Lags Behind
A study by Boomi and Forrester published in July 2026 underscores the urgency of clear rules. While 86 percent of surveyed companies have launched pilot projects with AI agents, only 34 percent of decision-makers actually trust those systems.
The trust gap widens sharply in organizations with uncontrolled AI deployment: the confidence rate drops to 22 percent. The financial fallout averages $2.1 million in extra costs from fines or customer losses.
An Expleo survey from June reveals how dependent workers have become. Ten percent of German employees say they can no longer do their jobs without AI. Nineteen percent cannot complete their tasks on time without technological help. Nearly one in three business owners describe workflows without AI as significantly more complicated.
Free AI Coding Tools Pose a Risk to Trade Secrets
Specialist lawyers are warning companies against using free AI coding assistants. Entering proprietary code into these systems can expose trade secrets. Legal protection for AI-generated code under copyright law also remains murky.
Their advice: businesses should negotiate specific contracts with vendors that guarantee confidentiality, filters for protected licenses, and indemnification clauses.
As you tighten data security around AI tools, don't overlook your physical safety documentation. A free Risk Assessment Toolkit with 41 templates and checklists helps you identify workplace hazards and record them properly — a core part of any compliance framework. Get the free Risk Assessment Toolkit
Google Ordered to Open Android to Rivals
The Digital Markets Act (DMA) is forcing Google to loosen its grip on Android. From July 2027, competing AI assistants must get full access to system functions — including voice commands and integration with Gmail or Maps. Starting January 2027, Google must also share search data with competitors.
Despite the regulatory squeeze, major tech players continue expanding. Oracle unveiled a studio for autonomous AI agents in July. Google rolled out new professional features for U.S. users in mid-July.
Industry observers such as Jonas Diezun, CEO of Beam AI, predict that autonomous multi-agent systems will soon handle tasks without human intervention. That, he says, will only intensify the debate over control and logging.
Disclaimer regarding our articles: No investment advice, no buy or sell recommendation. Information on prices, companies, and markets is provided without guarantee; changes are possible at any time. Stock market transactions can lead to substantial losses. Our articles are created and reviewed in whole or in part automatically with the support of AI.
