How is AI Affecting Talent Pipelines in Tech?

Learn to code was a rallying cry for technologists just a few years ago, and now it seems that advice has lost its power. In 2024, Nvidia CEO Jensen Huang said, 'It is our job to create computing technology such that nobody has to program.' 

But what happens if we rely too heavily on that computing technology and don't invest similarly in human expertise to use it effectively? 

It starts from the bottom.  

Automating routine tasks and freeing up senior staff to focus on more complex tasks is the goal. But that race for optimisation and the potential for revenue gains could compromise – and most likely already is – the future of technical expertise. 

For people starting their careers, AI is filling the roles they expected to fill. But it's about more than graduates getting jobs; there is potential for a talent pipeline collapse. Where does the practical experience that helps entry-level graduates gain critical expertise come from? 

 

Three recent perspectives offer different answers.  

• Stanford Digital Economy Lab points to a widening employment gap for young workers in AI-exposed occupations. (Stanford Digital Economy Lab, 2026) 

• The Federal Reserve Bank of New York urges a distinction between changing tasks and disappearing jobs. (Federal Reserve Bank of New York, 2026) 

• Anthropic’s internal research shows how work can shift towards directing and checking AI systems. (Anthropic, 2026) 

Together, they suggest that technology leaders need to think beyond immediate productivity and consider how their next generation of specialists will develop.  

 

Is AI already reducing entry-level opportunities? 

Employment among 22–25-year-olds in highly AI-exposed roles is around 19% lower than that among similarly aged workers in less-exposed sectors. 

The Stanford Digital Economy Lab used payroll data to identify differences between entry-level workers in tech and those in less AI-heavy sectors, also finding that experienced workers showed no comparable gap. 

That is not the same as saying AI has eliminated 19% of junior jobs, but it does point to a trend that, if accelerated, could drastically limit the number of specialists in a number of critical areas. 

There are limits to the findings. The study spans occupations beyond technology; its findings are descriptive rather than causal. Other labour-market forces may explain part of the divergence. Yet the concentration of declines in occupations where AI automates tasks, rather than complements workers, makes the pattern worth watching. 

 

Are jobs disappearing, or are employers asking for different skills? 

AI is changing hiring needs more clearly than it is driving widespread job elimination. 

In New York, firms reported few AI-driven layoffs, even as some anticipated reducing future recruitment. Other firms were increasing hiring for people with AI skills. (Federal Reserve Bank of New York, 2026) 

The useful distinction is between a job and the tasks within it. Automating a first draft, a routine query or part of an implementation process does not necessarily remove the need for the person responsible for the outcome. 

This does not contradict the warning about younger workers. Economy-wide employment can remain resilient while specific entry routes become harder to access. Equally, an occupation can survive while its expectations change substantially. 

For junior technology professionals, that makes the question less about whether an AI tool can produce an answer, and more about what they contribute around that answer: understanding the requirement, recognising errors, testing assumptions and explaining the result. Employers need to decide which of those capabilities they can reasonably expect on arrival, and which they must help people develop. 

 

What happens when AI becomes part of the team? 

AI agents and tools are increasingly common in teams, though their introduction constantly shifts the technical expertise and experience needed to monitor and manage them effectively. 

Anthropic’s latest research offers a company-level view of that transition. Its internal study reports a smaller share of junior technical hiring as AI-agent use expanded. It also finds that output rose as employees adopted new AI tools, while hiring priorities increasingly emphasised autonomy and judgement over implementation and specific technical expertise. (Anthropic, 2026) 

The implication is not that human expertise becomes unnecessary. Rather, the point at which it is needed shifts. More generated work creates more material to evaluate, coordinate and monitor. The company describes employees building tools and processes to oversee agent activity as human attention becomes a bottleneck. (Anthropic, 2026) 

That creates an awkward entry-level paradox. If organisations increasingly value independent judgement, but automate the work through which judgement is developed, newcomers may be expected to arrive with experience they have had fewer opportunities to gain. 

There is an opportunity here too. With appropriate supervision, AI could help an early-career professional explore alternatives, test ideas, and contribute to broader projects sooner. But access to faster output is not automatically access to deeper understanding. 

In a best-case scenario, AI agents develop junior talent better than before with computing power far beyond that available to human mentors. The worst-case scenario is a classic example of trying to run before you can walk. Junior talent might be thrown into positions, having completed AI training, but lacking the general critical-thinking ability to perform. 

 

How can technology employers protect the route to expertise? 

Adopting AI and developing talent are separate management challenges. One does not solve the other, and protecting the pipeline comes down to more than AI literacy.  

For project leads and technology managers, four questions offer a useful starting point: 

• Which tasks are being removed, and what did they teach? Map the learning previously gained through routine implementation, documentation and troubleshooting. Where that practice disappears, replace it with supervised exercises, paired work or structured reviews. 

• Does the entry-level role still have an entry point? Separate essential starting skills from capabilities the organisation will teach. A vacancy that requires independent technical judgement should explain how a newcomer will build it, rather than disguising an experienced role with a junior title. 

• Is AI use supported by meaningful feedback? Give newcomers opportunities to explain their reasoning, test outputs and discuss mistakes with experienced colleagues. Faster delivery should leave room for questions, not remove it. 

• What is happening to the future talent pipeline? Review junior recruitment, access to mentoring and progression alongside productivity measures. Short-term delivery gains deserve scrutiny if they come with a shrinking route into specialist work. 

None of this requires preserving repetitive work simply because it has always existed. The aim is to preserve its developmental value while finding better ways to deliver it. 

 

Will AI remove the first rung, or help rebuild it?

The answer is still up for debate.   

The evidence does not support a single verdict that AI is either taking entry-level technology jobs or creating an uncomplicated opportunity. The recent research describes different parts of the same challenge: a widening gap for young workers in exposed occupations, changing demand for skills, and a shift towards human oversight in an AI-intensive workplace.  

The question organisational leaders should be asking themselves is bigger than how much work a smaller team can deliver today? Instead, they should be asking: how do we secure the expertise capable of leveraging new technology tomorrow? 

AI may change the first rung of a technology career. Employers still have a choice about whether newcomers can reach it.  

If you're looking for expertise that understands the current challenges and is prepared to tackle the challenges of tomorrow, then schedule a call with us today. With an international network of specialists, we know what it takes to get a project over the line and how to prepare for the future.  

 

References

Athreya, K. B. (2026, August 5). AI’s impact on labor and hiring. Liberty Street Economics. https://libertystreeteconomics.newyorkfed.org/2026/08/ais-impact-on-labor-and-hiring/ 

Brynjolfsson, E., Chandar, B., & Chen, R. (2026, August). Canaries in the coal mine? Six facts about the recent employment effects of artificial intelligence [Working paper]. Stanford Digital Economy Lab. https://digitaleconomy.stanford.edu/publication/canaries-in-the-coal-mine-six-facts-about-the-recent-employment-effects-of-artificial-intelligence/ 

Wilmers, N., Catipon, R., Massenkoff, M., & McCrory, P. (2026, October 16). How employment is changing at Anthropic. Anthropic. https://www.anthropic.com/research/employment-at-anthropic