On December 13, 2023, a Workshop on AI Training in Catalonia, organized by CIDAI and AIR. The workshop brought together professionals from the academic, industrial and administrative sectors. This article summarizes the main conclusions of the debate.
The Catalan ecosystem in AI, sustained by knowledge generation and talent Catalonia is currently one of the most prominent European hubs in Artificial Intelligence. With a strong academic base of long tradition, the country’s universities and research and technological centers have 28 AI-intensive SGR research groups from 13 different entities, that lead the generation of knowledge internationally, and generate abundant talent. This talent, combined with the growing demand for professionals in the digital sector, and the growing establishment of technology companies or the creation of new ones start-ups “Deep tech” is the cornerstone of the ecosystem that makes AI a strategic sector for economic growth and improving people’s lives. According to the report “Artificial intelligence in Catalonia”[1], prepared by ACCIÓ in collaboration with the Secretariat of Digital Policies through CIDAI, are identified up to 488 companies working in AI in our country, of which 44% are start-ups, and 24 are spin-offs created by research groups. These companies charge more of 2,000 million euros (0.8% of Catalan GDP) and generate nearly 15,000 jobs. In Catalonia we can observe a phenomenon similar to that experienced in Quebec during the first decade of the century, with a turning point in 2013 with the creation of the laboratory MILA[2], where strong academic institutions with a great capacity to generate talent, associated with industry as an employer of this talent and engine of innovation, with the support of the government, make Quebec a hub globally recognized for research and innovation in AI. With the launch of the Catalonia.ai strategy[3], promoted by the Government of the Generalitat of Catalonia in 2020, the country has a roadmap that promotes AI focused on talent as an engine of growth, aiming for a model of technological sovereignty.
The training offer and demand for AI profiles in Catalonia Catalan universities have incorporated master’s and doctoral training programs focused on AI and data engineering, in addition to those already existing in various areas of ICT. These programs not only offer a solid theoretical foundation, but also include internships in companies and applied research projects. This allows students to develop important skills, such as machine learning, natural language processing and computer vision, which are highly valued in the job market. Catalan industry is experiencing a significant transformation with the adoption of AI to improve competitiveness in sectors such as health, automotive, tourism and logistics, among others. This demand has caused companies to actively seek professionals trained in this field. According to recent data, the AI job market in Catalonia is expected to grow exponentially in the coming years, with a notable increase in job offers.[4].
The current training offer is not sufficient to meet demand. The regulated technical training offer is limited to approximately 8,000 people/year trained, in contrast to the entire regulated offer which is +55,000 people/year. Universities should adapt their teaching offer to market demand, increasing the number of places in degrees in this field, or creating new master’s degrees. The needs are not only in volume but in content. There is a need to improve the offer to make it more flexible and adaptable to the evolution of technology itself. The involvement of companies in some undergraduate courses is noteworthy, but the need to increase the role of companies in curricular subjects (via external internships or final undergraduate/master’s theses) and the participation of these companies by proposing new challenges and participating, in general, in the definition of the training topic, adapting it to the future needs of the company is also evident. On the other hand, companies value the technical capabilities of professionals trained in academia, but they clearly see the need for more functional profiles, with complementary skills (soft skills) for know how to adapt AI models to business reality and the client context (it is necessary to develop project management, sales, communication, legal and ethical aspects, etc.) In certain cases, there is an over-demand for capabilities. No need for “superheroes” nor that all profiles are for university graduates. Job offers must be appropriate to needs.
Need for sectoral specialization and reskilling There is a need for continuous training, with sectoral specialization and reskilling, and also for AI training at other levels of public training provision, such as vocational training. The reskilling It is key to being able to meet the growing demand for talent from companies. Currently, AI has become a tool for competitive improvement in all economic sectors. The offer of micro-degrees or microcredentials by universities is a key factor in “lifelong learning”. The academy of the future must allow professionals from other specialties to be trained in AI to facilitate adoption in their sector. On the other hand, AI professionals must have tools to keep up to date, as well as to acquire complementary skills. Companies value that the training offer can be fast and agile and with diversity of formats for better adaptation to needs (intensive 1-year programs, postgraduates, micro-credentials, bootcamps, face-to-face, online and “in Company” courses…) Challenges The current supply of talent is limited, while the need for talent on the part of companies is increasing, to the point that some see their growth limited. At the same time, regulated technical training is limited in terms of maximum student capacity and lack of flexibility, due to high administrative management and insufficient economic resources for adequate updating and expansion to meet this growing demand. This raises a challenge in attracting and retaining talent both in universities and in the public sector. The acceleration in technological evolution around AI generates the challenge of adapting the training offer in AI to the reality demanded by the market. Additionally, there is a need to automate operational and business processes through the adoption of AI.
Catalonia is a country of SMEs, and there is a barrier to access to talent for small companies, which cannot compete with large ones in attracting this talent. It is therefore a challenge to introduce SMEs into the equation. It is a challenge to promote coordination and collaboration with the innovation ecosystem and business groups in the territory, such as PIMEC, CECOT and FOMENT to design instruments to bring and consolidate AI innovation to companies.
There is a challenge in moving from knowledge production to the generation of AI-based products. The research carried out in Catalonia does not have the same impact on innovation and the transfer of technologies generated to production. This situation causes a flight of talent and knowledge generators outside the territory, which also penalizes the local training ecosystem. Large international AI companies offer highly competitive contracts, but this puts the knowledge generated in the country at the service of third parties. We need to work to continue the integration of these companies in the territory, having them establish laboratories in Catalonia, and offering co-ownership instruments based on the transfer of talent. On the other hand, we must support entrepreneurship by offering facilities to young talent to create locally established start-ups. For this reason, it is necessary to promote mechanisms to capitalize on these initiatives.
Generative AI will impact the training and activity of more technical profiles, but these will continue to be necessary. The impact will above all involve a redesign of their activities around programming, since generative AI will be able to facilitate the performance and validation of the most routine activities, and will make it possible to devote greater time to tasks with greater added value.
[1] “Artificial Intelligence in Catalonia”, ACCIÓ Report, 2024. https://www.accio.gencat.cat/ca/serveis/banc-coneixement/cercador/BancConeixement/eic-la-intelligence-artificial-a-catalunya [2] MILA is the Montreal Institute for Learning Algorithms https://mila.quebec/en [3] Artificial Intelligence Strategy of Catalonia – Catalonia.AI. https://politiquesdigitals.gencat.cat/ca/economia/catalonia-ai/ [4] Digital Talent Overview 2024. https://mobileworldcapital.com/en/trends/digital-talent-overview-2024/
