Demand for people who can build and manage artificial intelligence (AI) has grown quickly. Lists of AI specialties usually read as career advice. This article looks at the same specialties from another side: what they ask of education, who is able to enter them, and what the growth of this kind of work means for society.
A field of many professions
AI is built by teams. The roles most often named are these.
- Machine learning engineers design and build systems that learn from data.
- Data scientists collect, clean and analyse the data those systems use.
- Computer vision engineers work on systems that interpret images and video.
- Natural language processing engineers work on systems that handle human language.
- Robotics engineers join AI to machines that act in the physical world.
- Software engineers, data analysts and research scientists build the surrounding systems, prepare data and develop new methods.
Beside them stand two roles defined by purpose more than by technique: specialists in AI ethics, and people who apply AI to public problems such as health care, climate and poverty.
What the field asks of education
The technical roles share a base: programming, most often in Python, together with statistics and linear algebra. That base is laid long before university. A pupil who leaves school without confidence in mathematics will find most of these doors hard to open later. The strength of AI education therefore depends on ordinary school teaching as much as on specialised programmes.
The field also crosses the usual lines between faculties. Language technology needs linguistics. Robotics needs mechanical and electrical engineering. Ethics work needs philosophy, law and the social sciences. Education systems that separate the sciences from the humanities at an early age prepare students poorly for work that needs both.
Who can take part
Much of the knowledge needed is openly available. Widely used tools are free, and courses and documentation can be read online. In principle a capable student anywhere can learn.
In practice there are barriers. Most teaching material is in English. Training large models needs computing power that few institutions can afford. Employers and research centres are concentrated in a small number of countries and cities. For small societies this raises a familiar concern: young people who gain these skills may have to leave in order to use them, unless remote work or local projects give them a reason to stay.
Ethics as part of the work
AI systems are now used in matters that affect people directly, so questions of fairness, transparency and accountability have become part of the work itself. Who is represented in the data? Can a decision be explained to the person it concerns? Who is responsible when the system is wrong?
These are not purely technical questions, and they cannot be left to engineers alone. They call for people who know the communities in which a system will be used. A tool built with data from one society may work badly in another, especially for a small language or a group that is thinly represented online.
AI and public problems
The use of AI for public benefit attracts much hope. It is sensible to treat that hope with care. Results depend on good data, on local knowledge and on institutions able to act on what a system reports. Technology can support such work. It cannot replace the people and organisations that carry it out.
A wider conversation
The variety of AI specialties shows that the field is not the business of programmers only. It needs teachers of mathematics, linguists, lawyers, social scientists and informed citizens. How a society educates its young people, and whether it includes its own languages and experience in the data from which machines learn, will shape what it gains from this technology. These matters deserve attention well outside the technology sector.
