AI competences: market situation and outlook

AI talent demand and supply

In the last 15 years the talent gap around artificial intelligence (AI) has widened, and continues to grow. Demand is so high for expert profiles (data engineers, data scientists, AI experts, and data technologies) that even now, despite an exponential increase in access to subjects and online content in the field, the increased educational need remains unmet. The cause is the democratization of the use and applications of AI – use at the level of home user, or as a professional user within a company is increasing.

Likewise, universities and educational centres (public and private) are not used to updating their agendas at the same pace as technology that long ago broke Moore's law, or that of technical publishers specialized in renewing annual editions. Kimball's The Data Warehouse Toolkit, one the main manual of business intelligence, stopped excelling once its annual printed editions became obsolete in a time when technology can sometimes substantially change within a three-month period, either in its components or in creation of new algorithms that presuppose leaps in predictive analytics, language processing, or artificial image generation.

Google, Amazon, Microsoft, Meta, Open AI, and universities such as Washington, Columbia, and MIT set the pace with monthly ‘white papers’ that challenge the state of the art in computer vision, language processing, character recognition, biometrical recognition, or quantum analytics and that also translate into a greater number of patents. Different types of LLMs (large language models) and agents represent a revolution in terms of the automation of tasks and work processes. However, Europe is still lagging behind in terms of research and patents. The number of papers published in Europe is half that in the US (and in China double that of the US, which appears to be looking to reach supremacy in AI before 2030). It must be considered that every 3.8 papers generate a patent. China clearly leads the global AI patent market in 2024, with 300,510 applications compared to 67,773 in the US and only 22,133 in Europe. That gives a sense of China's power, dominating 70%, according to mescomputing.com.

The development and promotion of AI is directly proportional to the investment in R&D&I and the development of the talent it produces – two fully-connected concepts. AI triggers a need for talent because we are talking about a technology that requires knowledge and experience for its application. Despite its rapid democratisation by ‘appification’ and ease of user experience, the consequences of misuse can be serious.

Therefore, the recommendation is to learn guidelines for use in the case of algorithms with medium and high risk, including certifications (like a "driver's license") that demonstrate that you have skill in the use and knowledge of the regulations (AI Act).

Trends

Trends will always be on the rise. Imagine that today the AI Expert is like the ‘shaman’ in the marketing profile of the 1980s. Actually, today every SME between 50 and 250 employees has at least one marketing profile. It would be necessary to have at least one AI expert for each senior marketing profile. That means that if there are 1,500,000 SMEs in Spain right now (according to the MINECO census) in case only 50% adopt AI technologies, 750,000 AI experts will be needed – in Spain alone – to be able to implement and apply AI in their companies.

According to Kai Fu Lee, an AI expert and former CEO of Google and Apple in China, the number of AI experts in the world has increased from 10,000 in 2020 to 100,000 in 2025 – that is, demand has multiplied 10-fold in five years.

Challenges

The challenges always have to do with minimising the impact of incorporating AI into business processes, how they affect current employment, and to what extent the substitution effect it may cause will broadly outweigh the net result.

It seems very clear that the employment generated by the implementation of AI can be much higher than that which it destroys – but the fact that it can cause such a change in employment patterns indicates the importance of implementing policies within companies that ensure the correct management of this change. To this end, the creation of AI Ethics departments is recommended to ensure the proper use of this technology, as well as the correct application of regulations.

The McKinsey Global Institute says that 14% of the workforce will be displaced by 2030 due to process automation. On the other hand, the World Economic Forum (WEF) has already warned of the elimination of 75 million jobs in the world by 2025, but 133 million new jobs could be created – that is, almost two would be created for each one destroyed.

However, Kai Fu Lee himself says that by 2035 40% of all jobs in the world will be done by some type of AI. AI will focus on replacing jobs that perform repetitive tasks or those that can be easily replaced by the following technologies:

• Customer and employee information service processes
• Business processes
• Surveillance and detection
• Autonomous driving
• Distribution, placement, and storage of goods
• Consultation and advisory processes
• Agriculture

Through 2035 we will observe a gradual process of total replacement of jobs. However, in other more complex tasks we will see a process of complementarity of AI and its incorporation into processes that help humans carry out their work more effectively. In many cases, AI will complement the human in their workplace or different functions.

Companies, governments, institutions, and civil society itself are responsible for ensuring that the process of transition and adaptation of people to new jobs is carried out in a balanced way, avoiding as far as possible that no group can be left out of the system.

Examples of new jobs

1. Agent coordinator: In charge of supervising AI agents that automate industrial processes.
2. Prompt engineer: Specialist in designing instructions to feed and train AI models.
3. Ethical and AI regulatory supervisor: Responsible for the ethical and regulatory application of AI models in accordance with the new AI regulation.
4. Data preparer: Person who gives coherence and consistency to the datasets so that they can be fed by AI.
5. Data platform engineer: Responsible for the design of shared data spaces from which AI is nourished.
6. Expert in AI: Developer in different languages of multilevel programming of AI models.

Conclusion

In conclusion, AI is a clear net generator of employment with a minimum ratio of 2 new positions for each one it replaces, and according to the latest study in this regard by the consulting firm E&Y, it will cause 84% of operational improvements in the processes applied. The real challenge is to try to retrain the replaced positions into AI profiles so that no one will be left out.

It is therefore advisable to start managing change so that people learn to live with machines. It is not that machines can replace people (which may happen in some processes), but that the machine makes the person more intelligent and empowers them to apply much more added value in their tasks, improve productivity, and better allocate their time to higher value-added tasks.