Entertainment

Correio newspaper | It seems like it, but it isn’t: fake videos, photos and audio manipulated by artificial intelligence are circulating on the internet

[ad_1]

MEDIA EDUCATION

It seems like it, but it isn’t: fake videos, photos and audio manipulated by artificial intelligence are circulating on the internet

See how to identify deepfake items that have been creating waves of scams, misinformation and risks for victims

Published on October 21, 2024 at 08:56

AI tools are increasingly advanced and the solution to combat deepfakes is prevention, in addition to developing a critical view of what is fake and what is real on the internet

AI tools are increasingly advanced and the solution to combat deepfakes is prevention, in addition to developing a critical view of what is fake and what is real on the internet Credit: Adobe Firefly

In this super digital era marked by selfies and short, fast-paced videos, many things seem real, but they are nothing more than deepfakes. The technique that, through artificial intelligence, manipulates images showing people saying what they never said or in acts they did not commit, is disseminated across the internet with increasing numbers of cases. Although there are still no specific statistics for Brazil, the international report Sumsub Identity Fraud Report 2023, which analyzed more than two million fraud attempts in 224 countries, indicates that Brazil presented the highest volume of sophisticated frauds, based on artificial intelligence, in Latin America. The increase in cases of deepfakes, between 2022 and 2023, was 830%, the highest in Latin America.

The report highlights that deepfakes appear first among the main frauds, followed by practices such as “money muling” – which is the recruitment of people to move illicit money into their bank accounts, facilitating money laundering – and falsification of documents identity. This data shows that, although the majority of known cases involve public figures, ordinary individuals are increasingly susceptible to becoming victims of deepfakes.

Proof of this is that two weeks ago, sponsored posts circulated across the country on Instagram, with videos in which presenters Angélica and Sabrina Sato appeared announcing a promotion on bags from a Rio fashion brand. To win the desired suitcase with tropical prints, the person would have to answer a survey, enter their details and, using a link, make a pix to pay for the gift to be sent. Pure scam. Many people fell because the videos were very convincing, but the presenters never recorded them.

In addition to promoting scams, losing money and exposing personal data, deepfakes can lead to many other serious consequences. In 2022, the author Glória Perez addressed the topic in the soap opera Travessia, in which the character Brisa (Lucy Alves) had a photo of herself used in a montage made with an AI application and posted online, causing her to be confused with a kidnapper. children. In the serial, the character was chased through the streets, beaten, arrested and even lost custody of her own son. A warning from the author about the dangers that lay ahead.

Another proof that the internet is not for amateurs was the situation faced by singer Taylor Swift, in January of this year. At the height of her career, the artist had false sexual images displayed online. One of the pieces of content made with AI had 47 million views in less than 24 hours, before being removed from X, formerly Twitter. And it doesn’t stop there, there are a series of videos, especially on YouTube, of fake romances between artists, and even famous singers’ voices in songs they never recorded, for example, BTS member Jungkook singing the song Die With Smile (Lady Gaga and Bruno Mars), which was recently launched when he was already away from artistic activities and serving the South Korean exercise.

Speaking of South Korea, Drama actresses, K-Pop idols and even women in the army did not escape deepfakes. In recent times, with the success of K-dramas and female music groups, countless fake sexual videos with their images have invaded Telegram groups. The so-called NTH Rooms – groups of men who post and consume pornographic images of women – have used fake digital montages. Korean culture researcher Thaís Midori published on her channel the information that 96% of deepfakes circulating on the internet today have pornographic objectives and 25% of them involve South Korean women.

The wave of Telegram groups reached schools, reaching young people, who manipulate images of colleagues, teachers, employees, sisters and mothers of other members within these groups. More than 200 schools have already been mapped, where women are now victims of these actions and blackmail. The army removed photographs of women serving in the military from its pages, and many protests are taking place to ask the government to act in defense of them and punish those who commit and/or disseminate crimes.

Recently in Brazil, presenter Sandra Annemberg also used social media to announce that she was a victim of deepfake. A video she had made for Globo Repórter was doctored with text that she never said in order to apply scams, using her credibility from years of television. “I feel outraged. Worst of all is that there are people falling for this fraud and this is very serious, very sad. People who believe in me think they are seeing something true and fall for false advertising. This is a crime”, she warned.

Raise a critical eye so as not to fall for a scam

Anyone with access to the internet can be a victim or consumer of deepfakes, especially young people who spend a lot of time online. Therefore, awareness-raising work in schools has been proposed and reinforced by the Instituto Palavra Aberta, through its EducaMídia project, which propagates the importance of media education.

According to Daniela Machado, education coordinator at Instituto Palavra Aberta, previously the basic level of care and attention was to observe details in these videos, such as lack of synchrony between speech and mouth movements or facial expressions, flaws at the edges of the images, inconsistent lighting and presence of shadows not consistent with the image shown.

The big issue is that the artificial intelligence tools that allow the generation of deepfakes are increasingly sophisticated and accessible, so many of these errors and flaws that were more noticeable to the naked eye are no longer existing, because the programs are enhanced, making identification more difficult.

“That’s where the decisive part of the critical look that we preach in media education comes in, to reflect on the context of that information, image, that video or that audio: is it bringing some revelation that could end someone’s reputation? Who is gaining from this content or, on the other hand, who is being harmed? Is there a public figure being exposed or someone you know? Does it really sound like something that person you know would do? What do you know about the source from which this content came? And those allegations that are being made, is there any evidence to prove that that was actually said? Can’t this information be found in sources, on already known and trustworthy journalistic websites? This more critical look will try to establish the context of the information there and this will be increasingly decisive, as those more technical flaws are corrected with the advancement of this technology”, teaches Daniela Machado.

She also highlights that many platforms have been working to make information increasingly clear and people are already thinking about creating ways to make the use of AI more transparent, for example, placing a seal or watermark on the information generated by AI. When in doubt, he recommends, the best option is always not to go over what you are not sure about.

OAB remains attentive to the scenario

Not only celebrities are exposed to the risks of deepfakes, but ordinary individuals are increasingly more likely to be victims of these crimes, due to the ease of access to artificial intelligence technologies and the wide availability of personal images on social networks, increasing this risk. . Because of this, the Permanent Committee on Technology and Information of the Bar Association – Seccional Bahia has been working to combat this situation. The lawyer, professor specializing in Cybercrimes and President of the Permanent Committee on Technology and Information at OAB Bahia, Tamíride Monteiro, highlights that the “Commission is aware of this reality and has sought partnerships with academic institutions and international organizations to deepen understanding of the scenario national and regional. We recognize the importance of collecting accurate data to support effective public policies and promote the updating of the legal framework, aiming to face the challenges posed by this emerging technology.”

To avoid falling for deepfakes, the best way is prevention, as explained by Victor Habib Lantyer, lawyer, professor, author and researcher in Digital Law and Artificial Intelligence and Artificial Intelligence coordinator of the Permanent Commission for Technology and Information at OAB Bahia. “It is recommended to limit the exposure of personal information on the internet, avoiding excessive sharing of photos and videos on public platforms. Keeping privacy settings restricted on social networks and being careful about the permissions granted to applications are also important measures”, he highlights.

He also emphasizes that, within the scope of digital security, it is advisable to use two-factor authentication in online accounts and be aware of possible phishing attempts that seek to collect personal data. “Educating yourself and those around you about the risks and signs of deepfakes can contribute to a culture of greater vigilance and collective protection,” he says.

In case you are a victim of deepfake, the following steps are recommended:

Collect evidence: Document all possible forms of evidence, including source files, metadata, conversation logs, and any interactions related to the content. This evidence will be crucial in investigations and legal proceedings;

Register a police report: Look for a police station specializing in cybercrimes, such as the Computer Crimes Repression Police Station. Formal registration is essential so that the competent authorities can act;

Seek legal assistance: Consult a lawyer specializing in digital law for specific guidance. The professional will be able to assist in taking appropriate legal measures, such as civil and criminal actions, in addition to guiding you about your rights;

Request content removal: Contact the internet service providers and social media platforms where the deepfake was posted. According to the Marco Civil da Internet (Law No. 12,965/2014), providers must remove content upon court order, but many have internal policies that allow voluntary removal in cases of rights violations.

What the law provides for internet crimes

Brazilian legislation, despite not having a specific rule on deepfakes, offers legal mechanisms to punish harmful acts resulting from this practice. Crimes against honor (slander, defamation and insult), provided for in articles 138 to 140 of the Penal Code, can be applied when the deepfake causes damage to the victim’s reputation or dignity;

The Carolina Dieckmann Law (Law No. 12,737/2012) typifies crimes related to the invasion of computer devices, and may be applied if the obtaining of images or videos used in deepfake occurred through unauthorized access to the victim’s devices;

The General Personal Data Protection Law (LGPD) (Law No. 13,709/2018) establishes guidelines for the appropriate processing of personal data. The misuse of images and personal information in deepfakes may constitute a violation of the LGPD, subjecting those responsible to administrative sanctions and requiring compensation for damages;

In cases involving minors, the Child and Adolescent Statute (ECA) provides for severe punishments for the production, distribution and possession of pornographic content involving children and adolescents, including digitally manipulated materials.

[ad_2]

Source link

Michelle Williams

I'm Michelle Williams, an enthusiastic author specializing in captivating entertainment content on Rwcglobally.com. With a passion for storytelling and a keen eye for the latest trends, I aim to engage readers with compelling narratives that reflect the dynamic landscape of the entertainment industry. Join me on Rwcglobally.com to explore the world of film, television, music, and more, as we uncover the stories that define contemporary culture.

Post Comment