During my 17+ years at Amazon and AWS, I have been a passionate advocate for investing in early career talent. Ever since the advent of generative AI in the mainstream however, I hear a steady stream of predictions about AI replacing humans, especially early career talent. If true, this would render investing into early career obsolete. I dug into the data and don’t think that’s right. In this post, I share what the evidence actually shows, how I think leaders should frame the AI question, and the principles I use to build early career talent programs that hold up in any environment.

What is happening?

Generative AI has been triggering the imaginations of business and technology leaders since the launch of ChatGPT in 2022. The hype about evolving capabilities of generative AI systems is reflected in regular headlines about entire job families being eliminated. A Stanford reference study ‘Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence’ from November 2025 identified six facts about recent labor market effects of artificial intelligence, including ‘substantial employment declines for early career workers in occupations most exposed to AI, such as software development and customer support’, but also ‘entry-level employment has declined in applications of AI that automate work, with muted effects for those that augment it’. The study found a real 16% employment decline for early career work that AI promises to automate.

AWS CEO Matt Garman took a different view: He called out the idea to replace all junior developers with AI as one of the dumbest things he ever heard. He pointed out that early career talent is the least expensive headcount and that they are most likely to lean into AI tools. Most importantly, he asked: what happens 10 years from now if you stop feeding your talent pipeline? After all, he started out as an intern at Amazon back in 2006, and made it to AWS CEO less than 20 years later. He also argued that AI tools can guide early career talent, making them more productive and accelerating their career growth. He emphasized that skills such as critical reasoning, creativity, and a learning mindset are becoming more important in our age of fast technology advancement.

That said, not every workforce reduction making headlines is AI-driven. Big Tech’s recent massive layoffs, including Amazon’s reduction of 16.000 employees, are better explained as a correction to the hiring surge before the post-COVID financial crisis than by AI replacing workers. What changed since 2022 is not just generative AI, but the entire macro-economic climate. Without real economic growth, companies are reluctant to invest in human talent regardless of what technology can do.

Two AI belief systems

Two schools of thought dominate the discourse on the impact of AI on human workforce: those who see AI as a means to reduce cost of labor by replacing humans, and those who see the potential of AI to augment human capabilities to not only accomplish more, but to tackle work they could not do before. This divide isn’t new. It mirrors how leaders have been approaching early career talent: those who see it as a cost-efficient way to perform repetitive but necessary work, and those who see it as an investment to train the next generation of leaders.

These two approaches to AI are not mutually exclusive. If work can be fully automated by an AI (or other system) that objectively performs as well or better than a human, we should challenge why humans should continue to do this work. However, we tend to overestimate the capabilities of current AI systems and underestimate the complexity and cost of building AI-powered systems that can reliably remove humans from the loop.

Invest in the future

A world where AI replaces the majority of human jobs is neither desirable, nor inevitable. We should continuously invest in human talent to achieve better outcomes for businesses and society with the increasingly more powerful tools available. In an environment of rapid change it is more valuable to focus on the invariants - the things that do not change - rather than the noise of change. Change will keep accelerating, but the need for human talent and leadership is durable. Here are the principles I keep coming back to when investing in early career talent.

Today’s early career talent is tomorrow’s leader

Investing in early career talent just because ’they are cheap’, or because ’they are most likely to lean into AI tools’ is not a sound strategy. You invest into the career growth of individuals with the expectation that they will thrive and become tomorrow’s leaders. Effective leadership requires a solid understanding of how a business works, which is why early career talent should be embedded in the business and do ‘real work’.

Integrate early career talent into your business

At scale, it may be tempting to formalize a training curriculum for consistency, but this comes at a cost: lack of exposure to how your business actually works. Instead, integrate them in your existing teams, and provide mentorship, coaching, and the opportunity to actively contribute. Above all, emphasize ownership - the courage to take responsibility for delivering real outcomes. Curiosity, critical thinking, and a learning mindset follow naturally from that.

Career development for mid-senior talent

To create a win-win situation, use your investment into early career talent also as an investment in your mid-senior talent. By integrating early career talent in your team, you create the opportunity for other team members to expand their skill set. For example, I have given ownership of recruiting and managing new interns to employees whose career growth plan included the desire to explore management. Others took on the responsibility to be mentors, which contributed to their own body of work to be promoted to the next level. There are many ways to identify such win-win situations. What they all have in common is that the team welcoming the early career talent has skin in the game to make them successful. However, this also means that each team has capacity limits in the number of early career talent they can support. As a rule of thumb, you need at least two experienced employees per early career talent they can coach.

Do the impossible

When integrating early career talent in your team, you have the choice of doing ‘more of the same’, or doing ’things you could not do without them’. I recommend running time-bound experiments exploring new approaches to solve known business challenges. Such experiments need single-threaded focus for a few months to deliver a prototype or to test a hypothesis. Unless your team has extra capacity (does that ever happen?), such experiments are impossible to conduct thoroughly, but they may be a perfect opportunity for an internship or ramp-up project.

Conclusion

If you are considering cutting your early career program because of AI or budget pressure, ask yourself this question: what does your talent pipeline look like in 10 years? The current macro-economic headwinds are real, but they are temporary. AI will change how work is being done, but it will not eliminate the need for human talent. Invest in early career talent today to benefit from the resulting compounding interest. Embed them in real work where they learn your business and seize the opportunity to develop your mid-senior team’s leadership skills. Finally, use their capacity to run the experiments you would otherwise never do.

In my career, I had the privilege to work with incredibly talented people. My investment in early career talent paid off with interest, as I have learned at least as much from them as they learned from me. I got to see them go from intern to managing large teams, build products used by thousands of people, speak at international conferences, and thrive in their careers. Investing in early career talent means shaping people’s lives. That’s a legacy no AI can replicate.