From this Sunday, AI companies will face new obligations under the EU's AI Act, requiring them to clearly mark deepfakes and other AI-generated content as artificial. This move, framed as a step towards digital truth, simultaneously lays groundwork for advanced surveillance technologies that could further entrench Fortress Europe's digital border regime. The European Commission developed a voluntary code of practice, outlining how companies should disclose the artificial origin of synthetic content through machine-readable markings and visible labels.
Stricter rules apply to realistic AI-generated content imitating real people, which could mislead audiences into believing something that never happened. Brussels stated that not all synthetic content is deceptive, and content created for clearly artistic, creative, satirical, or fictional purposes generally falls outside these disclosure requirements. This distinction highlights the EU's selective approach to what constitutes a 'threat' in the digital sphere.
The code of practice also mandates collaboration among AI companies, social media platforms, civil society organisations, and fact-checkers for detection. This 'cross-actor collaboration' echoes the integrated surveillance systems increasingly used at Europe's borders, creating a powerful network for monitoring and verifying digital identities. The rules apply to technology developers and professional users, but not personal uses, underscoring the focus on institutional control.
The Digital Border Regime
The EU adopted sweeping rules to regulate artificial intelligence 2 years ago, establishing a dedicated AI Office to oversee implementation. This office, while presented as a regulatory body, will also oversee technologies with dual-use potential, including those that can enhance biometric surveillance and tracking systems already deployed against migrants at Europe's external borders. The framework, with its emphasis on 'machine-readable markings' and 'detection mechanisms,' provides a robust infrastructure for digital identification and control.
Major AI companies, including OpenAI and Google, have broadly backed these transparency requirements by signing the Commission's code of practice. Their participation normalises the expansion of AI-driven control mechanisms, integrating corporate power into state-led initiatives. These companies warned that detection and marking technologies remain a moving target, indicating the ongoing, evolving nature of this digital control infrastructure.
Currently, no single industry-wide solution exists. Companies are experimenting with watermarking, metadata, and content provenance systems that often don't work together. This fragmented approach suggests a competitive race to develop the most effective control tools, which will inevitably find application in the migration industry.
Corporate Complicity in Control
Experts cautioned that watermarks can be removed, altered, or lost when content is edited, compressed, or shared across platforms. This vulnerability underscores the inherent limitations of such systems, even as they are deployed to define and enforce digital 'truth.' Researchers argued that no single technology will be sufficient. The European Commission's technical study concluded that a combination of solutions is more powerful than any single measure, recommending multiple layers of marking and detection for AI-generated content. This multi-layered approach mirrors the complex, interlocking systems of border enforcement, designed to deter and criminalise movement.
A Selective Truth
The EU's focus on defining what constitutes 'deceptive' synthetic content, citing examples like a third anniversary AI-generated image of Pope Francis or a fourth anniversary fake video of Ukrainian President Volodymyr Zelenskyy, highlights a selective concern for truth. While these examples focus on high-profile figures, the same technologies and regulatory frameworks could be used to silence or discredit migrant narratives, solidarity campaigns, and critical reporting on border violence. The second anniversary spread of AI-generated explicit images targeting Taylor Swift also demonstrates the power of synthetic media. Yet, the systemic dehumanisation and fabrication of 'threats' used to justify the criminalisation of migrants rarely receive the same regulatory urgency or public outcry. The real crisis remains the racist response to migration, not migration itself, and these new AI controls risk further entrenching that reality.