On February 2, 2025, the first compliance deadline of the EU AI Act went into effect, ushering in a new era of accountability for artificial intelligence technologies. Among its many provisions, the Act takes a firm stance on deepfakes by requiring transparency and disclosure from businesses that create or deploy such technologies.

For enterprises and organizations across all industries, this regulation is a clear signal that deepfakes are no longer a futuristic concern, but a very real threat impacting businesses today. Recent data reveals that deepfake incidents worldwide increased by 245% year-over-year (Sumsub), with 25.9% of executives reporting one or more deepfake incidents targeting their financial and accounting data in just the past year (Deloitte). As a company focused on securing critical communication channels against deepfake impersonations, we see this as both an opportunity and a challenge for businesses to take proactive steps in safeguarding their operations.

The Labeling Mandate

The EU AI Act mandates that any AI-generated or manipulated content must be clearly labelled as such. This includes the use of watermarks or other technical markers to ensure that audiences can easily identify deepfake content. While watermarking and provenance solutions are valuable tools in the fight against deepfakes, they face significant limitations. According to recent research, adversaries have demonstrated ways to remove or tamper with watermarks (FS-ISAC), and these solutions only work when content creators actively participate in marking their content as synthetic. This leaves organizations vulnerable to bad actors who intentionally circumvent or ignore these requirements.

(In contrast, inference-based deepfake detection — which analyzes content directly for signs of manipulation — offers a more robust and platform-agnostic approach. Unlike provenance solutions, inference detection works regardless of whether the content creator chooses to participate, making it particularly effective against malicious actors who deliberately avoid transparency measures.)

No Longer Hypothetical Risks

The risks posed by deepfakes are not hypothetical — they are happening now, and the EU AI Act clearly indicates as such. According to recent studies, 51.6% of executives expect an increase in deepfake attacks in the coming year (Deloitte). We've seen real-world cases where deepfake technology has been used for executive impersonation scams, disinformation campaigns, and reputational attacks. These types of attacks don't just threaten financial stability — they erode trust among employees, customers, and stakeholders.

Without the ability to detect deepfakes in real time, enterprises and governments face substantial risks of reputational damage, breach of confidential data, and theft of assets. The EU AI Act's focus on transparency is a step in the right direction, but it also places significant responsibility on businesses to adopt tools and strategies to detect and mitigate these threats before they cause harm.

At Reality Defender, we believe this regulatory shift is a wake-up call for businesses to take deepfakes seriously. With 70% of global decision makers now viewing deepfakes as a meaningful threat to their businesses, organizations must invest in proven, multimodal detection solutions that can identify manipulated content in real time. It is no longer enough to react after the damage is done; businesses need proactive systems that protect their critical communication channels from malicious actors.

The fight against deepfakes is not just about compliance — it is about preserving trust in an increasingly digital world. As deepfake technology continues to evolve, organizations must act now to implement robust detection solutions that not only meet regulatory requirements but serve as the last line of defense against an emerging wave of AI-powered fraud that threatens to undermine the very foundation of digital trust.