The AI Skills Tug-of-War

As artificial intelligence (AI) has moved from experimentation to expectation, both employers and employees agree that AI skills matter, and that more training, capabilities-building, and skills development are needed to meet AI goals.

The more uncomfortable question is: Who’s responsible for developing those skills?

Research we recently conducted at Emergn shows a significant disconnect between employers and employees over who’s responsible for AI training. The vast majority of employees (81% percent) say it’s the responsibility of employers to upskill staff. By contrast, 83 percent of CEOs believe employees should take ownership of upskilling themselves.

This is much more than a philosophical disagreement. As expectations grow around AI to drive revenue and business growth, a lack of skills, capabilities, and training are already causing delays in transformation initiatives and negatively impacting productivity, mental well-being in staff, and career progression.

For the training and learning industry, this moment represents both a challenge and an opportunity to help organizations move beyond the “who’s responsible?” debate and toward shared, practical models for AI capability-building that work.

Making a Case

For employees, the logic is straightforward: Building and honing new skills in service of meeting corporate ambitions should fall on employers to facilitate. This is especially true when it impacts not just workflows and decision-making but also can influence performance reviews, career growth, and job security.

Indeed, our research shows the push to adopt new AI tools and platforms is predominantly coming from employers, with 89 percent of CEOs saying they intend to drive revenue from new AI solutions in the coming year. The desire to see employees equip themselves with AI capabilities likely reflects concerns about cost, speed of change, and the perceived risk of investing in upskilling human capital who may leverage those learnings to secure a new role elsewhere.

The result is a standoff between employees waiting for guidance and investment, and employers waiting for initiative and self-directed learning. In the meantime, AI adoption efforts have stalled or produced uneven results—which ultimately hinders an organization’s ability to remain competitive.

The Intelligent Delusion

As we’ve seen with other emerging technologies such as cloud computing, Internet of Things (IoT), and machine learning, success stems less from knowing how to use a tool and more from understanding the problems your organization is trying to solve with it.

Using AI to achieve goals such as cutting costs or reducing headcount is short-sighted and limited in scale. In contrast, embedding AI to create new revenue streams, improve data-driven decision-making, and reduce redundant or tedious work so employees can focus on strategic initiatives offers far greater potential for long-term business growth.

This is a common trap for business leaders that we call the “intelligent delusion.” This is where organizational leadership sees the adoption of new technologies as a sign of advancement when, in reality, deeper levels of organizational change around workflows, processes, and even how decisions are made are required for true transformation.

This is the difference between simply using AI as a tool and adopting an AI-first mindset. Organizations that fail to adopt an AI-first mindset risk overestimating the promise of AI while underestimating the human element required for long-term success. Bridging gaps in talent, skills, and organizational capabilities is a critical component of creating and embedding an AI-first mindset to achieve scalable growth.

A Shared Responsibility Model

To successfully bridge these gaps, the best path forward lies in shared ownership where employers and employees each play distinct, complementary roles.

What employers can own:

• Strategic direction: Clarifying where and why AI matters, its alignment with business goals, and which problems teams are expected to solve.
• Foundational enablement: Providing baseline AI literacy, ethical guidance, and learning pathways for employees.
• Learning infrastructure: Providing time, tools, coaching, and safe environments in which to experiment and apply new skills.
• Organizational context: Teaching how AI fits into products, services, and value streams—not just how tools work.

What employees can own:

• Active participation: Engaging with provided learning, experimenting responsibly, and applying insights to real work.
• Continuous learning mindset: Staying curious as tools and techniques evolve.
• Feedback and insight: Sharing what works, what doesn’t, and where additional support is needed.
• Career alignment: Connecting new capabilities to personal growth and long-term employability.

One of the best ways to operationalize a shared approach is through a product-centric approach where the focus is placed on delivering continuous value through quality and innovation. This approach is becoming increasingly prevalent as it empowers teams to embrace experimentation, adopt cross-functional collaboration, and better manage uncertainty.

From a training perspective, this means:

• Teaching teams to frame AI initiatives around customer or user outcomes, not technology hype.
• Developing skills in problem discovery, experimentation, and measurement—areas where AI can amplify impact.
• Encouraging cross-functional collaboration, since AI-driven products and services rarely sit neatly within one role or department.

By positioning AI training through a product-centric lens, organizations make learning immediately relevant and actionable. Employees see how AI connects to their day-to-day decisions, while leaders gain a clearer view into business value.

Practical Considerations for Leaders

For leaders looking to equip their teams for AI-driven growth, several practical steps stand out:

1. Anchor training to real work. Avoid generic AI courses. Tie learning to current products, services, or internal processes so employees can apply skills immediately.
2. Define “good” clearly. Be explicit about what effective AI use looks like in different roles —from ethical considerations to decision-making boundaries.
3. Create learning pathways, not events. One-off workshops are insufficient. Build progressive pathways that combine foundational knowledge, hands-on practice, and ongoing coaching.
4. Reward learning behaviors. Recognize experimentation, knowledge sharing, and continuous improvement—not just short-term results.
5. Balance autonomy with support. Encourage self-directed learning, but back it with time, budget, and leadership endorsement.
6. Measure outcomes, not attendance. Track improvements in cycle time, quality, customer outcomes, and employee confidence. These are metrics that matter to both the business and the learner.

From Tug-of-War to Traction

Despite differing viewpoints on who’s responsible for AI training and upskilling, the greatest threat to both employers and employees lies in inaction. By meeting in the middle, combining employer-led enablement with employee-driven growth and grounding training in product-centric thinking, organizations can turn the AI skills tug-of-war into shared traction. For the training industry, the opportunity is clear: Help both sides move from disagreement to durable capability, and from technology adoption to meaningful, human-centered progress.