- Open administration
The AOC organizes a workshop to advance the digitalization of applications for school lunch subsidies
The belief that Artificial Intelligence (AI) can tackle the complexities of almost any problem is attractive and has gained many followers recently, following the emergence of generative AI, which does things with an ease that was unimaginable until recently.
However, this perspective overlooks a crucial point: AI technology is simply a tool, not a solution in itself. While AI can dramatically improve efficiency, it cannot solve the human and organizational complexities that constitute the fundamental problem of public administration.
As Jim Collins emphasizes in his extraordinary book “Good to Great”, successful organizations are not dazzled by trendy technological trends. They first clearly define their mission, understand the essential factors of the challenges and create teams with the right talent, using technology as a catalyst for transformation. This principle applies to the Administration.
Despite significant technological advances over the past 20 years, including the internet, mobile phones and social media, the foundations of government structures and cultures have remained and have proven remarkably resistant to change. Unfortunately, too often we have digitised old, outdated processes, rather than rethinking the public sector to deliver truly innovative and citizen-centric services. We must be aware that unless accompanied by deep organisational transformation, AI can add another layer of complexity without creating substantial public value.
Many challenges in the public sector, such as inefficient regulation, hyper-bureaucratic structures, disincentive organizational cultures, lack of impact evaluation, and difficulties in attracting talent, are fundamentally human problems.
AI is undoubtedly an extraordinarily powerful tool for improving efficiency in certain areas, but it has inherent limitations. For example, generative AI solutions do not understand their actions, they lack an understanding of the meaning behind words: they are not intelligent in any deep sense, nor do they have empathy or emotions (at least with current models).
Generative AI is an advanced tool that predicts the next word in a sentence based on the vast amount of text it has processed, which includes almost the entire internet. Essentially, it operates like a very sophisticated calculator, guessing the most likely word rather than actually understanding its meaning.
Therefore, AI is incapable of promoting public values, negotiating compromises, building consensus, or generating trust. Nor can it provide inspiring leadership, engage with a collective project, manage strategic change, facilitate cultural transformation, or design empathetic, citizen-centered services. All of these responsibilities are inherently the work of people.
In short, we should not delegate our essential human responsibilities to machines. We need to empower public servants with the right tools, including AI, to promote more efficient, kinder, and people-centered governments.
References: