Meanwhile, AI is also advancing relentlessly in the field of political communication. It's being used to formulate arguments, articulate hordes of social media posts, and even improve the tone of voice and diction of political leaders in the clips of their speeches that are broadcast. This is how political marketing expert Isaac Hernández explains. It was already being used in the 2023 general election campaign, and continues to be used now: from fake news to montages with fabricated speeches about Pedro Sánchez, to Donald Trump's use of it, portraying himself as the Pope or at a fictional tourist resort in Gaza. "It's a constant communication machine. And, in light of that, the media plays a fundamental role as a reference for what's true," the neuromarketing expert claims.
Catching corrupt officials, detecting fraud, or helping you with your personal income tax: the new uses of AI
Big data allows for the generation of alerts on suspicious bids by scanning public procurement portals.
BarcelonaThe artificial intelligence (AI) revolution has reached the administration. The public sector has been incorporating massive data processing tools for a few years now, but the acceleration of these technologies is now opening up new horizons: for example, to catch corrupt officials, detect tax fraud, help taxpayers file their tax returns, or fine those who don't have an environmental sticker on their cars by scanning their license plate. Some of these mechanisms are already part of the daily work of the State Tax Administration Agency, Social Security, and some city councils. Others are just beginning to gain momentum. This is an issue that is gaining importance in current politics, following scandals such as the Cerdán and Montoro cases, but also the Spanish State's failure to comply with the anti-corruption commitments demanded by the Council of Europe.
But how can AI help detect corruption cases such as the one affecting the PSOE and the Ministry of Transport of José Luis Ábalos? According to Júlia Miralles, associate professor of political science at UPF, the systems of big data They already allow public procurement portals to be scanned to generate alerts and review suspicious contracts. They do this in a much less cumbersome manner than a traditional audit, despite the added difficulty posed by the fact that, at present, there is no full interconnection between this and other data held by the State on all those involved. However, from a technical point of view, it is not a difficult or expensive technology to implement, according to expert sources consulted by this newspaper.
This system allows for the detection of calls red flags to identify overcharges that may be concealing commissions, whether for masks or public works. In this field, some initiatives are already underway: the European Commission has created a tool called Arachne to detect irregularities in European funds, and the Catalan and Valencian Anti-Fraud agencies have also taken steps in this area. During the appearance of President Salvador Illa Regarding the Cerdán case, the Commons asked the Government to roll up its sleeves to strengthen these tools in the Generalitat.
The most innovative cases
Beyond Europe, there are even more innovative cases. With support from Microsoft, the Inter-American Development Bank has created a platform, Investment Map, which uses AI to increase transparency in public investments. In Costa Rica, authorities have joined the big data with the so-called Internet of Things (IoT) and 5G. The result has been a camera system through which citizens can follow the execution of public works live. The system can detect who enters the construction site, but also whether the volume of concrete, cement, or metal used actually corresponds to the amount tendered. "But it's still very experimental," notes Miralles, a specialist in public integrity and fraud prevention.
In China, a system was implemented a couple of years ago that processed all types of data on public officials to prevent corrupt behavior. Despite the positive results, it was discontinued, and the government did not provide further details. In this sense, the expert warns of the need to debate what the limits of these tools should be in order to respect people's rights (particularly the right to privacy).
However, not all corruption is detectable with these systems, adds Ismael Peña-López, professor of law and political science at the UOC. Cases such as the Montoro scandal, in which he allegedly used a law firm to solicit money from companies in exchange for tailoring laws, are key. "The administration is very uncorrupted because it's very well-organized," notes the expert, who emphasizes that, to make a real leap forward in fraud prevention, it is necessary to advance the professionalization of public sector managers. It is one of the measures that the Government has implemented.
The great challenges
The implementation of these technologies in the administration faces two major challenges, according to the experts consulted: first, data protection legislation. Second, administrative processes, which are often slow and do not evolve at the pace that AI has in the last two years. Furthermore, the efficiency of these tools often depends on the data available to the administration and the planning it carries out for its public policies. And this is where there is great room for improvement, Peña López emphasizes. who has also studied this issue in depthIn other words, for AI to work, it first needs to have good databases and know what it will be used for, keeping in mind the biases it may have toward vulnerable groups. If these obstacles are overcome, in a few years, progress could be made in preventing natural disasters, major droughts, or fires.
To stay ahead of these technologies, the Spanish government, the Catalan government, and local governments have sought out technology companies familiar with the latest trends in the AI sector. Lluís Vicente, head of consulting and innovation at the Catalan firm Quantion, highlights, first of all, the growing use of chats. bot on public sector websites: they serve both to help a person apply for a grant and to provide customer service to citizens. This AI specialist also detects a growing interest in knowing how citizen data can be protected when deploying these technologies. "We're talking about sensitive information. Sometimes people seek us out because we provide technical, but also ethical, advice," he explains. Cybersecurity experts also come into play here. José de la Cruz, technical director of the multinational Trend Micro in the Iberian Peninsula, maintains that AI, despite being a secure technology, tests the capabilities of administrations: "The bad guys are increasingly more efficient. And the line between good and bad is sometimes very thin," he adds.
