The EU AI Act marks its second anniversary of coming into force this August, as a seminal piece of wide-ranging emerging technology legislation. It is the first of its kind in the world, but recently debate has shifted from the legislation itself to its implementation.
With pushbacks from the private sector, and concerns over meeting stringent regulation on time, key elements of the landmark regulation are delayed. Examining this, industry leaders argue that extending deadlines should not be seen as a weakening of Europe’s AI ambitions.
In fact, many believe it presents an opportunity for organisations to build more robust governance frameworks. Four experts from Vanta, PagerDuty, Pipedrive and Coursera highlight the delays provide time to invest in AI literacy, and embed responsible AI practices that will support both compliance and innovation in the years ahead.
Tackling the AI compliance question
Khush Kashyap, Senior Director of Governance, Risk and Compliance, Vanta said: “The ‘watering down’ framing misses what deadlines actually do to compliance programmes. When teams are racing the clock, they document what they have, they don’t build what’s needed. You end up with security theatre: the appearance of governance without the infrastructure to back it up.
“The delay creates real space to do this properly. Most product teams still haven’t had a serious conversation about which of their AI systems qualify as high-risk under the Act’s risk tiering, what that means for their development process, or where human oversight is genuinely required versus retrofitted after the fact. AI asset inventories with real provenance trails, impact assessments that aren’tlegal box-ticking, governance that’s operationalised rather than just documented is an enormous amount of work.
“The EU AI Act remains a global benchmark for compliance, and this delay will ultimately make for better products. It’s a practical shift, and it encourages the yardstick for success to be about product quality rather than legal compliance.”
The importance of aligning on policy
Joao Freitas, Chief AI Officer, PagerDuty said: “Delay to EU AI Act implementation timeline needs to be viewed as an opportunity. Organisations within Europe have the chance to strengthen alignment between policy ambition and operational reality.
“AI integration speeds on. Organisations across Europe are already integrating AI into critical workflows and digital operations – what they need now is practical guidance that is consistent and implementable at scale. The EU AI Act has a key place in this picture; clear governance frameworks, transparent standards, and collaboration between regulators and industry will be essential to building long-term trust in AI adoption.
“At PagerDuty, we believe responsible AI governance and transparency need to be embedded into how organisations design, deploy, and monitor AI systems from day one. With some of these timelines phased later through the EU AI Act’s implementation delays, this change is not an excuse to pause work on AI management and best practices. Emerging technologies should not be treated as part of an isolated, standalone compliance exercise. A phased implementation gives space for experimentation and the time to operationalise those controls effectively. This will help ensure the EU AI Act supports both trust and continued innovation across the European digital economy.”
Strengthening transparency
Sean Evers, VP of Sales and Partner, Pipedrive said: “AI does help sales teams work faster, but there is concern and even pushback in its rapid growth.
“Greater adoption will depend on transparency, control and trust being shown to both the internal and external users of systems – users within the business and their public customers. Automations reduce admin and free people up for better customer conversations. If it is both fit for purpose and aligns with all stakeholders’ expectations, it helps sales teams work smarter to focus on building loyalty and exceptional customer experiences.
“Businesses will struggle if they use AI to automate any existing poor habits at scale. Generic outreach, opaque lead scoring or AI-generated messages sent without review will immediately damage trust. It’s wise to allow an AI assistant to suggest which deals need attention or draft a follow-up, but have the salesperson apply their human empathy, context and judgement before acting.
“Legislation such as the EU AI Act can help all business functions, including sales innovation, to focus on better practices. Its enduring value will be to encourage the use of AI in ways that are transparent, controlled and recognisably human, such that the business finds efficiencies, internal users gain time and focus, and customers find an improved experience.”
Building out AI literacy and skills
Anthony Salcito, General Manager, Coursera said: “One unacknowledged benefit of extended EU AI Act implementation timelines is the increased degree of opportunity it offers institutions to ensure that their workforces are equipped with the critical skills required to adhere to all of the Act’s legal and operational requirements. Companies must remember that compliance with the EU AI Act creates an urgent skills imperative, with recent Coursera research indicating that 88% of tech leaders fear that their AI transformation goals will not succeed without greater investment in talent development – and 72% expect all employees, including those in non-technical roles, to understand AI’s implications for their own work.”[Text Wrapping Break][Text Wrapping Break]“There is still time to alleviate these fears. On Coursera, we see that learners can reach intermediate proficiency in AI & Big Data skills with approximately 80 hours of study — and beginner proficiency in only 30. The European Parliament has given European companies an expanded window to nurture an AI-literate European workforce, and one that is more than sufficient to ensure that the EU labour market is equipped to innovate responsibly, safely, and competitively by August 2028.”
What’s over the horizon for the EU AI Act?
While opinions differ on the pace of regulation, there is consensus that the success of the EU AI Act will ultimately depend on how effectively organisations operationalise its requirements rather than simply meeting deadlines.
Experts broadly agree that any additional implementation time should be used to strengthen governance, improve transparency, develop workforce skills, and ensure AI systems are deployed responsibly.
As organisations get their governance procedures in order, several of the EU AI Act’s risk-mitigation measures are already in force, including bans on unacceptable-risk AI systems, even as some later implementation deadlines are being reconsidered.
With degrees of higher risk are already being mitigated, the pause button is being pushed on some of the finer points, including mandatory risk management systems, data governance requirements and technical documentation. If used well, the delay could help deliver not only stronger compliance but also more trustworthy and resilient AI across Europe’s digital economy.
