Technology makes us more productive. When a new technology emerges, it allows us to delegate tasks to machines that historically required human expertise.

These technology transformations have an impact on what jobs will be in demand in the future. Today, organizations and their employees are determining how AI will impact their current roles and long-term career paths.

AI experts have differing opinions on how AI will shape career paths. Vinod Khosla, the billionaire AI investor, believes that AI will eventually make expertise — and most jobs — irrelevant. At the other end of the spectrum, some AI leaders believe that highly specialized expertise will be extremely valuable when paired with AI skills. I call employees who are able to combine subject matter expertise with fundamental AI abilities “AI+X employees.”

Based on my experience teaching AI at Stanford University and as founder of an AI skills intelligence platform, I believe the true answer lies somewhere in the middle. Students should absolutely get a broad education, focusing on the durable skills that will matter throughout their career. But in a world being reshaped by AI, expertise will be more valuable than ever before. AI+X employees will be the leaders and visionaries who drive us forward.

Understanding AI models

Why do I think expertise is going to be so important? It has to do with the way multimodal AI models are developed and trained.

Multimodal models are trained on diverse content — primarily generated by humans — which shapes their understanding of the world. While synthetic data can be generated, cutting-edge scientific content will still rely heavily on human expertise until we reach a certain level of artificial general intelligence (AGI).

For example, when given a sentence and asked to predict the next word, a multimodal AI model draws on its training content to make the prediction. We can think about the content used to train an AI model as a circle. Huge amounts of human knowledge can be found within the circle: everything from math courses to books in Japanese. The model will be able to make connections and extrapolate using the content in the circle: it will be able to teach math in Japanese without being trained specifically on a Japanese math book. The model will master the content within the circle; where it gets more difficult is at the edge of the circle — the cutting edge areas that require extreme levels of expertise.

Many of the world’s most technical subjects — things like material science, cybersecurity, propulsion, structural engineering — aren’t covered extensively in existing, publicly available content. If there is detailed content available, it’s often private property buried inside a company’s database — or within the brains of a few experts.

As the workforce develops and begins using AI models as a tool, we will see humans doing less and less work within the circle. Employees will instead begin using their expertise to extend the boundaries of the circle, increasing what becomes common knowledge and then researching further. The perishable skills found on the edge of the circle will drive innovation; the durable skills within the circle will sustain that innovation.

What is an “AI + X” employee?

To be an AI+X employee, you need the X — an area of defined expertise that can be extended and amplified using AI. AI itself is now a durable skill comprising domains such as data science, analytics, machine learning, and statistics. Employees must reach a standard level of proficiency to apply AI in their everyday work. For experts, AI will be a constant requirement even as the more specific skills in their industry emerge, evolve, and become obsolete.

AI+X employees are important to organizations because they allow companies to build new solutions and capitalize on new opportunities more quickly. It may take a few months to train a chemical engineer to use AI tools, but if we look at it from the opposite direction, it would take several years for an AI practitioner to reach the same level of knowledge in chemical engineering. A subject matter expert who can add AI to their skillset will be a better long-term asset to their organization than a generalist who is highly proficient in AI.

Subject matter experts also play an essential role in giving a stamp of approval to AI systems. We need these leaders to evaluate AI-powered solutions and confirm whether they meet the highest standards for their industry. Without AI+X experts, we have no way of knowing if an AI model has gone off course.

I’ve seen the potential of AI+X employees in the classes I teach at Stanford: the students that don’t come from a computer science background consistently come up with fascinating, innovative applications for the technology. I had a group of students with experience in oil and gas, and they developed a project using AI to conduct predictive maintenance for a drilling system. If a drill is several kilometers beneath the Earth’s surface, a broken drill is catastrophic — being able to predict damage is massively valuable.

Another group of students came from electrical engineering. They developed a camera that can be ingested by patients which will go through your digestive system, take pictures, and predict if the patients have certain pathologies. This solution couldn’t be developed by a computer scientist — it required electrical engineering and medical expertise.

Find Problems, Not Technologies

Inventive solutions like AI predictive drill maintenance and a diagnostic AI camera could only be imagined and built by people with deep levels of expertise and a passion for the problem they’re trying to solve. This is why AI+X people will be so essential to the next generation of technology solutions: they’re committed to solving a problem.

All employees should be developing their AI skills, but AI isn’t a problem to be solved — it’s a technology. Companies and employees exist to solve problems. This is another way to think about the X in AI+X — it’s not just an area of expertise, it’s a passion to solve a particular problem. 

Today’s founders and organizations need to make sense of the AI frenzy, and they need a strategy to find and develop their AI+X employees. The best approach they can take is to avoid getting too lost in the tools and technologies and to stay focused on their reason for being — the problem that they exist to solve.