Is everyone ready for the new EU regulations going into effect on 2 Aug?
The EU AI Act's transparency requirements mandate that providers and deployers of AI systems clearly disclose when users are interacting with AI or viewing AI-generated content.
This is every time an AI is interacting directly with a natural person it must explicitly disclose it is an AI in a clear and timely manner.
When interacting directly with a natural person the AI must explicitly disclose it is an AI in a clear and timely manner. For agentic voice assistants (like automated receptionists, customer service bots, or outbound sales agents), this means:
Upfront Audible Warning: The voice agent must announce itself as an AI system at the very beginning of the call, before any sensitive data is collected.
No Buried Disclosures: Relying on disclosures in Terms & Conditions, hidden URLs, or purely text-based machine-readable logs is not compliant for direct interactions.
Technical Labelling: Agentic voice architectures must embed auditable watermarking and metadata indicating synthetic audio generation.
Language: The disclosure must be delivered in the language the conversation is being conducted in.
Enforcement for these rules began on August 2, 2026, backed by the European Commission's Code of Practice on Transparency of AI-Generated Content. These transparency regulations come into strict enforcement across the European Union, with penalties reaching up to €35 million or 7% of global annual turnover.
Furthermore, because live voice calls process sensitive data like names, phone numbers, and conversation transcripts, GDPR compliance is also fully required. My article about these other compliance issues to consider can be found at https://lnkd.in/d6mTEuvu