Unlike what is often assumed, the retirement of this generation does not imply an inevitable reduction in the size of the State. Official projections from the Civil Service estimate a flow of 104,789 new hires through 2035, which will more than compensate for losses and even increase the total number of staff. It is a qualitative matter.
“Unlike what is often assumed, the retirement of this generation does not imply an inevitable reduction in the size of the State”
The 2026 Public Employment Offer, with 27,232 vacancies in the AGE and more than 1,700 dedicated to information technologies (representing a 42% increase from the previous year), reflects an initial attempt to realign the profiles. Yet, will these new hires serve to endow the State with its own technological capacities? Or will the renewal be limited to contracting the management of systems developed outside the public sphere?
From Administrative Automation to Legal Guarantees
As analyzed in the report Governing in the AI Era, prepared by the Future Government Center of NTT DATA, current technology does not stop at digitizing procedures, but acts at the core of administrative activity. These are tasks ranging from interpreting documentation and classifying files to supporting the decisions made by administrations. In Finland, the Social Security Institution Kela uses artificial intelligence to classify the documentation of benefit applications. Thanks to this technology, it saves each year a workload equivalent to what 38 full-time employees would perform.
In the Spanish legal framework, automation encounters very clear limits. For example, Article 41 of Law 40/2015, of October 1, on the Legal Regime of the Public Sector, requires that any automated administrative action clearly identify the competent body, establish the system’s design and auditing parameters, and guarantee citizens’ rights to supervision and recourse. That is, AI can expedite the preparation of drafts or the detection of anomalies, but the weighing of public values and legal responsibility inevitably remains in human judgment.
Technical Sovereignty versus Private Intermediation
The main strategic risk for the Administration is not the slowness in adopting technology, which is also emphasized, but the loss of autonomy vis-à-vis private providers. This has been seen in recent weeks, with the government’s renewed push to dispose of services contracted to the American Palantir. In this regard, the NTT DATA report warns of two intertwined dangers: contractual dependence and, above all, the displacement of the direct relationship between the State and citizenship. If citizens begin to interact with public services through virtual assistants and private AI platforms, these corporations come to condition access to rights, the prioritization of information and the flow of relational data. In such a scenario, the Administration risks being relegated to a secondary role, limited to funding and sustaining public services, while the private sector controls the relationship with citizens, the data it generates and user trust.
“The main strategic risk for the Administration is not the slowness in adopting technology, as is often emphasized, but the loss of autonomy in relation to private providers”
Preserving the capacity for public decision-making requires retaining internal competencies to audit models, demand contract portability, and govern data as common infrastructure. But, looking beyond our borders, the truth is that this sovereignty is far from the norm. According to World Bank data (GovTech Maturity Index), only 23% of the evaluated economies effectively and jointly use a governance architecture, an interoperability framework and a shared data exchange platform. In Spain, initiatives such as the IA360 Plan, the ALIA strategy for language models in official languages or the creation of the Spanish Agency for the Supervision of Artificial Intelligence (AESIA) aim to consolidate a homegrown model. Nevertheless, without strategic public procurement that breaks supplier concentration and strengthens State autonomy, regulatory frameworks may remain merely on paper.
2035 Scenarios and the Future of Institutional Capabilities
The window of opportunity to define the State model is narrow. The application of European obligations on high-risk artificial intelligence between 2027 and 2028 sets a near horizon for decision-making. The same NTT DATA report outlines four future scenarios for 2035 according to the capacity for public foresight and control of the relationship with society. In the “Administration that Leads,” the State leverages the generational turnover to integrate technology and public employment under a coherent strategy, preserving human judgment and algorithmic transparency. At the opposite end, in the “Intermediated” or “Substituted” Administration, the lack of internal capacities makes institutions captive clients of corporate platforms, which means losing not only data control but the very ability to design public policies with their own criteria.
“The lack of internal capacities turns institutions into captive clients of corporate platforms”
The coincidence of the retirement of 90,000 public employees in the AGE and the maturation of artificial intelligence does not force the State toward an irreparable decline in relevance, but it does require an explicit political choice. If the 104,000 positions planned for the next decade are limited to covering obsolete structures or outsourcing the intelligence of services to technology companies, Spain will consolidate a dependent and opaque Administration for citizens. Conversely, if the replacement of staff is used to recruit and train in-house digital talent, audit the algorithmic systems and ensure that public interaction remains direct and transparent, the State will preserve the legitimacy and autonomy needed to govern in the digital era.
