With the EU AI Act enforcement postponed to December 2027, enterprises worldwide face an urgent need to ramp up compliance efforts amid a complex regulatory landscape. Recent feedback from CIOs highlights gaps in readiness, particularly in AI content watermarking and scope assessments, while organizations cautiously reconsider reliance on frontier AI models.
- Only 3% of enterprises have completed AI content watermarking.
- 36% of CIOs haven’t assessed EU AI Act applicability yet.
- Enterprises are cautious about frontier AI model dependency.
What happened
The European Union has postponed the enforcement of its AI Act to December 2027, providing enterprises additional time to prepare. Despite this extension, many CIOs and digital leaders are not advancing quickly enough toward compliance. Research from the diginomica network reveals that only 35% of surveyed digital heads have begun tracking or labeling AI-generated content, while a minuscule 3% have fully completed the watermarking process. Additionally, more than a third of companies have yet to determine whether the Act applies to their operations.
This lack of preparedness comes against a backdrop where digital leadership is still committed to staff training on AI literacy. Around 38% of respondents are moving towards compliance in training, and 13% consider themselves fully compliant. The focus of the EU AI Act on ensuring that people understand and manage AI’s effects is viewed as a positive regulatory angle by many enterprise leaders. Alongside the EU’s regulation, companies must also monitor burgeoning AI-related rules across different jurisdictions such as US states and Latin America.
Why it matters
The EU AI Act addresses fundamental tensions between AI innovation, regulation, and data sovereignty. It places people at the core of AI governance by emphasizing transparency, accountability, and mandatory training. The imminent regulatory framework aims to deter unchecked deployment of AI technologies by requiring AI content watermarking and compliance checks, which are essential to avoid penalties and build user trust.
At the same time, the debate around dependency on frontier AI models underscores risks related to technical reliance on external AI providers and large hyperscalers. Enterprises and governments are considering local AI models and sovereign data centers to mitigate risks and foster genuine innovation. Failure to balance regulatory compliance with strategic model choices could stifle innovation or enforce a status quo where AI adoption is limited to copying existing solutions.
What to watch next
Enterprises must accelerate their compliance efforts over the coming 18 months, especially concerning AI content watermarking and comprehensive assessments of AI use cases under the EU AI Act. CIOs will need to increase coordination with legal and compliance teams and invest in AI literacy programs for staff to meet regulatory expectations and maintain operational agility.
Parallel to regulatory preparation, organizations will monitor international AI governance trends and reevaluate their reliance on frontier AI models. The dialogue on building proprietary or localized AI capabilities versus depending on global providers will intensify, with implications for innovation strategies and geopolitical positioning. Observers should watch for how enterprises balance regulatory demands, operational needs, and technological innovation as the deadline approaches.