From 2 August 2026, new transparency rules under Article 50 of the EU AI Act apply to certain AI systems. Accordingly, end users must be informed when they are interacting with AI. AI-generated or manipulated content must also be labelled in a format that digital systems can recognize. Additional information may also be required when using deepfakes or biometric data.
AI devices need clear privacy and safety mechanisms
Products such as Meta’s AI companion Muse Charm and its AI glasses show why transparency must be considered right from the start. These devices can interact with their surroundings and may record people nearby. Therefore, visible lights or other signals can indicate when a device is recording. However, users have, at times, already found ways to disable or bypass these systems. Accordingly, transparency features should be made to be reliable and difficult to misuse. Data protection is key in all AI-endeavours. Cameras, microphones, biometric scans and other smart devices must meet all relevant AI Act and data protection requirements.
Compliance goes even beyond data protection
As AI products are also part of physical devices they will have to meet certain other requirements as well. These include electrical product safety rules, producer responsibility obligations and WEEE registration. Packaging may also require registration, system participation and correct volume reporting. It is essential to review all relevant requirements before bringing AI devices to market.
Need help with AI compliance, product registration or packaging obligations? trade-e-bility can help your business identify the relevant requirements and prepare for them.
