The use of social networks by people who are already very polarized could compromise rational democratic debate, leading to a world of disinformation, fake news and the manipulation of images and characters by artificial intelligence (AI). The warning comes from Spanish sociologist and professor Manuel Castells, one of the speakers at the International Seminar on Democracy and New Technologies, which the Senate will be promoting between March 25 and 27.
Castells says that the misuse of technology is worsening at a time when humanity is facing major problems. According to the professor, public debate has become a series of manipulations that have a negative effect on democracy. In an interview with TV Senado, Castells says that technology can increase the crisis of democracy. He cites the example of the elections in the United States and Brazil, when polarization and the use of social networks were very evident.
The Spanish professor points out that technologies are not the cause of the problem, but they "amplify it enormously". For Castells, the origin of the democratic crisis is the lack of trust in politicians and political institutions. He points to the regulation of social networks and the use of artificial intelligence as a possible way out.
According to Castells, three quarters of citizens don't trust their politicians or democratic political institutions. He also reports that 77% of those surveyed are in favor of democracy. However, the professor ponders, people consider that "the democracy they live in is not a democracy". So, since people already think that politicians are unreliable, that democratic politics is not what they live in in their country, social networks with all this disinformation greatly amplify the crisis of legitimacy and, increasingly, people react emotionally. They don't believe either the media or what they see on the networks, because everyone believes their part of what they see on the networks.
- I believe that technology in itself does not cause social and political effects. It's how technology is used, by whom and for whom. So what is happening is that as political processes are decided, above all in people's minds, the technological capacity to make technological communication processes based on fictitious visions of reality, (becomes) deformed according to political and social interests.
Mindful of Castells' warning, senators have presented several bills dealing with the use of artificial intelligence. For Senator Styvenson Valentim (Podemos-RN), author of the first bill presented in the Senate on the subject (PL 5.051/2019), artificial intelligence can bring major productivity gains in industry and service provision. However, he points out, artificial intelligence cannot be adopted without minimum regulation. Styvenson is also the author of the bill that establishes the National Artificial Intelligence Policy, with the aim of stimulating the formation of a favorable environment for the development of AI technologies (PL 5.691/2019).
Styvenson's matters are being processed together with several other proposals in the Temporary Internal Commission on Artificial Intelligence (CTIA). The CTIA was created in August last year to analyze, within 120 days, the preliminary draft presented in December 2022 by the Commission of Jurists on Artificial Intelligence. Based on the jurists' text, the president of the Senate, Rodrigo Pacheco, presented Bill 2.338/2023, which regulates the use of AI. The text creates rules for intelligence systems to be made available in Brazil, establishing the rights of people affected by their operation. It also defines criteria for the use of these systems by public authorities, providing for punishments for any violations of the law and giving the executive branch the prerogative to decide which body will oversee and regulate the sector. At the end of 2023, the CTIA was authorized to function until May 23rd of this year. The committee is chaired by Senator Carlos Viana (Podemos-MG) and its rapporteur is Senator Eduardo Gomes (PL-TO).
For Pacheco, the development and popularization of AI technologies has revolutionized various areas of human activity. He says that "forecasts indicate that artificial intelligence will bring about even more profound economic and social changes in the near future". According to the president, the legislative proposals presented indicate that the Senate recognizes the relevance of the issue. Pacheco also notes that his bill "seeks to reconcile, in legal discipline, the protection of fundamental rights and freedoms, the valorization of work and the dignity of the human person and the technological innovation represented by AI".
Senator Randolfe Rodrigues (Sem partido-AP) is the first signatory of a proposal to amend the Constitution to include the protection of mental integrity and algorithmic transparency among fundamental rights and guarantees (PEC 29/2023). Randolfe argues that there is a consensus that artificial intelligence mechanisms are promoting major propositional changes in everyday life, speeding up procedures, uniting wills and helping in the search for solutions, especially on virtual search platforms. However, the senator notes, the advances point to a "well-founded and real concern about the ethical and normative limits to be observed by neurotechnology, also bringing up increasingly frequent discussions and studies about digital addiction, especially in children and adolescents". The PEC is awaiting the appointment of a rapporteur in the Constitution and Justice Committee (CCJ).
Some bills seek to crack down on crimes committed using artificial intelligence. Two of them are by Senator Chico Rodrigues (PSB-RR). Bill 145/2024 prohibits the use, in an advertising message, of artificial intelligence to generate the image or voice of a living or deceased person without their clear and unequivocal consent, or that of the holder of the image right. The advertisement must also state that it has used artificial intelligence. If these requirements are not met, the advertiser could be punished for misleading advertising. The penalty is imprisonment from three months to one year and a fine. The television channel, website or other means of broadcasting can also be penalized if they do not stop broadcasting within three days of notification of the infringement, which can be made by government inspection bodies or even by the holder of the offended party's image rights.
Bill 146/2024 amends the Penal Code (Decree-Law 2,848 of 1940) to increase the penalty for crimes against honor committed on social networks that use deepfakes with artificial intelligence. Anyone who uses this type of tool to produce a fake video or image could face a fivefold increase in the penalty for a crime against honor - such as slander or libel. This could mean up to three years in prison for libel, or ten years for slander, as well as a fine. In the author's opinion, the current text of the law does not curb the seriousness of the offense that deepfakes can generate.
- The harmful potential of creating such material and disseminating it is far greater and more devastating than it was in the past, when we created our Penal Code. For this reason, I propose that the penalties for creating material using artificial intelligence be increased fivefold. But the damage is caused not only by the person who creates the material, but also by the person who disseminates or distributes it - Rodrigues explained to Agência Senado.
Senator Jader Barbalho (MDB-PA) has presented a bill that criminalizes the crime of unauthorized image manipulation (PL 623/2024). Under the bill, the manipulation of a photograph or video without the victim's authorization, with or without the use of technological resources, with the intention of producing an image of nudity, a sexual or libidinous act of an intimate nature, could result in a prison sentence of one to two years, as well as a fine. The penalty can be doubled if the crime is committed against a minor, and even tripled if the material is disseminated on social networks.
Another bill, by Senator Jorge Kajuru (PSB-GO), increases the penalty for anyone who uses artificial intelligence to edit an intimate photograph or video. Under Bill 5.722/2023, the penalty will be one to two years in prison. Today, the penalty is six months to one year. The fine provision remains. In the same vein, Senator Weverton (PDT-CE) also presented a bill to criminalize the creation, use and dissemination, without the victim's consent, of fraudulent content containing sex, nudity or pornography, including involving children and adolescents (PL 5.721/2023). Weverton's, Kajuru's and Jáder's bills are being analyzed by the Communications and Digital Law Committee (CCDD).
A bill has also reached the Senate that increases the penalty for crimes of psychological violence against women by 50% when committed with the use of AI (PL 370/2024). As a result, the penalty could be up to nine months' imprisonment. The bill was authored by Congresswoman Jandira Feghali (PcdoB-RJ) and is awaiting distribution to the committees.
In addition to acting as rapporteur for the CTIA, Senator Eduardo Gomes is the author, together with Congressman Eduardo Bismarck (PDT-CE), of a bill that creates the Permanent Joint Committee on Data Protection, Artificial Intelligence and Cybersecurity (CMCiber - PRN 01/2024). The senator argues that the importance of creating the commission lies in "the pressing need to regulate, regulate and monitor issues related to data protection, artificial intelligence and cybersecurity, issues that have gained increasing relevance in the national and global context". The proposal still needs to be voted on by the National Congress.
Senator Angelo Coronel (PSD-BA) presented a bill to create the Joint Parliamentary Front for Artificial Intelligence, Information Technology, Cybersecurity and Combating Disinformation (PRS 93/2023). The name could be shortened to Parliamentary Front for Digital Security. Both deputies and senators can join the front. Coronel says that the phenomena of artificial intelligence, cyber security and disinformation itself present themselves as challenges for the world. Hence the importance of the front. The matter is awaiting the appointment of a rapporteur in the CCDD.
Set up at the end of March 2022, the commission promoted 12 panels and 1 international seminar to discuss the regulation of artificial intelligence, as well as discussions on topics such as concepts, understanding and classification of artificial intelligence; impacts of artificial intelligence; rights and duties; accountability, governance and oversight. The commission had 18 members and its rapporteur was Laura Schertel Mendes, a professor of civil law. Justice Ricardo Villas Bôas Cueva, from the Superior Court of Justice (STJ), chaired the committee. The commission's final work, with more than 900 pages, was delivered to President Pacheco at the end of 2022.
The final text presented to Pacheco contains principles, rules, guidelines and foundations to regulate the development and application of artificial intelligence in Brazil. There are three central pillars: guaranteeing the rights of people affected by the system; gradation of the level of risks; and provision for governance measures applied to companies that supply or operate the artificial intelligence system.
Concern about the misuse of technology also affects elections. A recent resolution by the Superior Electoral Court (TSE) for this year's municipal elections characterized deepfakes as audio or video content, digitally generated or manipulated by artificial intelligence, to "create, replace or alter the image or voice of a living, deceased or fictitious person" (Resolution 23.732/2024). The use of this resource is prohibited. In addition, there will be restrictions on the use of robots (computer programs) in contact with voters. It will also be mandatory to insert a warning about the use of AI in electoral propaganda. With this validation, society will be able to be informed, with a label on the screen, whether the image broadcast is true or not.
These are some of the proposals and measures being discussed in the Senate to deal with the challenges and impacts of artificial intelligence on society and democracy. The regulation and responsible use of AI are essential issues to guarantee the protection of individual rights and the integrity of public debate in an increasingly digitized world.