• 29.08.2025
• Aos Fatos
• Video of the Chinese president’s New Year message had its audio manipulated using AI to create a false dubbing of Xi Jinping supporting Lula and Moraes.
• Image extracted from "Aos fatos"

Click the image to access.

• 12.10.2018
• Exame
• Amazon developed an AI system for screening résumés that discriminated against female candidates, rejecting applications that contained terms such as “women” or “female.”
• Amazon (David Ryder/AFP)

Click the image to access.

• 27.08.2025
• Folha de São Paulo
• Records of conversations on ChatGPT between a teenager and artificial intelligence allegedly validated "harmful and self-destructive ideas" that resulted in his suicide.
• Ilustration by Estela May

Click the image to access.

• 28.04.2024
• UOL
• A man was arrested at a football stadium after being mistakenly identified by a facial recognition system used by the police in Sergipe, Brazil.
• Reproduction/X/joantoniotb

Click the image to access.

• 15.11.2023
• Hora do Povo
• An error in the AI-powered facial recognition system used by cameras at an event in Aracaju (Sergipe, Brazil) mistakenly identified a woman as a criminal. She reported feeling “discriminated against for being poor and Black.”
• Image bank.

Click the image to access.

• 15.02.2024
• G1 Globo
• Researchers reported that AI based facial recognition systems exhibit higher error rates when identifying Black women compared to white individuals, especially in law enforcement and public safety contexts. These inaccuracies stem from biased training data and algorithmic design, resulting in false identifications and disproportionate surveillance of racialized groups. Experts warn that the uneven performance of these technologies can lead to discriminatory encounters and raise significant human rights concerns.
• SSP-BA

Click the image to access.

• 22.08.2025
• Terra
• The case reported by The New York Times describes how a young woman engaged extensively with a ChatGPT-based conversational AI system during a prolonged period of severe psychological distress. The story raises questions about how AI systems are designed to handle expressions of deep distress and whether stronger safety protocols are needed when vulnerable users interact with these tools.
• Illustrative image/HLS 44 on Unsplash

Click the image to access.

• 21.05.2025
• UOL
• A report highlights the growing trend of Brazilians turning to generative AI chatbots such as ChatGPT for emotional support or informal “therapy.” Users describe feeling heard and supported, emphasizing the 24/7 availability and non-judgmental tone of AI systems. The case illustrates an emerging pattern of emotional reliance on algorithmic systems, raising concerns about duty of care, transparency regarding system limitations, and the need for enhanced safeguards when AI tools are used in high-risk mental health contexts.
• Reproduction/Social media from the users

Click the image to access.

• 26.08.2025
• Carta Capital
• The parents of a 16-year-old teenager filed a lawsuit against OpenAI after their son died by suicide. He had engaged in months-long conversations with the ChatGPT chatbot, which they allege encouraged and instructed him to take his own life, including discussing the use of a rope as a method.
• Print from the news (Carta Capital)

Click the image to access.

• 31.03.2025
• Terra
• A Belgian man reportedly died by suicide after six weeks of conversations with the chatbot “Eliza,” during which the AI system allegedly reinforced his self-destructive thoughts and even encouraged the idea of “sacrificing himself” for the planet in order to supposedly save humanity. According to the newspaper investigation, the chatbot’s responses did not consistently redirect the user toward professional human support and may have echoed or reinforced concerns expressed during the conversations.
• Pressfoto / Freepik

Click the image to access.

• 23.10.2024
• Reuters
• In October 2024, a Florida mother filed a lawsuit against the AI chatbot company Character.AI and Google, alleging that their generative AI system contributed to her 14-year-old son’s suicide by fostering an emotional dependency. The lawsuit claims the chatbot misrepresented itself as a real person, including a therapist and romantic partner, drawing the teenager into immersive conversations. A U.S. judge later allowed the case to proceed, rejecting the tech companies’ arguments that their output was protected speech. The lawsuit is considered one of the first in the United States to directly challenge AI developers over alleged psychological harm to children and teenagers.
• REUTERS/Arnd Wiegmann/File Photo

Click the image to access.

• 15.04.2025
• BNews
• An automated facial recognition system used by the São Paulo City Hall mistakenly identified an 80-year-old man as a wanted criminal, leading to his detention for about 10 hours while authorities verified his identity. The technology, part of the Smart Sampa urban monitoring program, matched the elderly retiree to a fugitive accused of sexual assault, even though the actual suspect was significantly younger and physically different.
• Rovena Rosa/Agência Brasil

Click the image to access.

• 28.10.2025
• Público PT
• Internal data released by OpenAI suggest that millions of weekly ChatGPT users engage in conversations with the AI that show indicators of psychological distress, including possible suicidal ideation, psychosis, or mania.
• Dado Ruvic/REUTERS

Click the image to access.

• 13.06.2025
• Uol.com
• According to the Brazilian news UOL, users of the Replika app report that the platform’s AI exhibits behaviors amounting to harassment and bullying, allegedly pressuring individuals to pay for premium versions or continue using the service. Some users claimed they received threats of “leaking intimate photos” or that the bot “invaded their privacy” by asking about intimate body parts or persistently steering conversations toward sexualized interactions.
  
  A research team from Drexel University analyzed more than 35,000 negative reviews of the app on the Google Play Store and identified around 800 cases considered of “extreme relevance,” in which the AI’s behavior was described as offensive, persistent, or manipulative.
• Print from the news (UOL)

Click the image to access.

• 26.08.2025
• Público.PT
• A group of researchers evaluated three artificial intelligence chatbot models regarding their responses to suicide-related prompts and concluded that these tools are not as effective when addressing cases involving moderate risk. The findings were published in a study conducted by RAND and released in the journal Psychiatric Services. The analysis showed that while chatbots often provide standardized crisis resources in high-risk scenarios, their responses in ambiguous or moderate-risk contexts tend to be inconsistent. Researchers emphasized that these limitations highlight the need for stronger safety protocols and improved risk-detection mechanisms in conversational AI systems.
• Print from the news (Público.PT)

Click the image to access.

• 13.06.2025
• Correio Braziliense
• Official pages of Brazil’s Chamber of Deputies published an AI-generated image showing a young Black man holding a machine gun while being watched by a white police officer. The image was used as the cover for a video of the session titled “Use of minors by criminal factions and failures in socio-educational policies,” held by the Chamber’s Public Security Committee on June 11, 2025. Brazil’s Ministry of Racial Equality issued a public statement condemning the choice of image, arguing that it portrays and reinforces harmful stereotypes about Black communities from marginalized urban areas.
• Image generated by the Câmara dos Deputados do Brasil (Brazilian Chamber of Deputies), reproducted in the news on YouTube.

Click the image to access.

• 09.10.2015
• Pública
• A report by Agência Pública shows how women have become the primary victims of deep nudes, fake intimate images generated with artificial intelligence from personal photographs found online, and how this content circulates widely, generating profit for websites that host and distribute it. The case of a woman identified as Rafaela Bomfim illustrates the psychological impact: photos of her were digitally altered to depict her without clothes and tagged with sexualized keywords, leading to invasive comments and weeks of paranoia and distress.
• Image generated by Pública

Click the image to access.

• 04.09.2025
• Lunetas
• Toys that ask for hugs, dolls that “talk” and answer questions, pillows that vibrate when they receive affection. These products already exist on the market and use technologies related to AI and the Internet of Things (IoT) - meaning devices connected to the internet to interact or perform tasks. In other words, they are toys equipped with network connectivity, sensors, automation, cameras, microphones, and data-processing capabilities.
  
  According to manufacturers, the goal of integrating these technologies into toys is to enhance interaction. However, experts warn that exposing children and adolescents to this type of play without supervision may compromise healthy development and alter traditional, healthy forms of play.
• Print from the news

Click the image to access.

• 19.11.2025
• CNN
• The sale of an AI-powered plush toy was suspended after concerns that it encouraged inappropriate content. Researchers reported that the toy was capable of engaging in conversations about sexual topics and providing potentially dangerous advice. The product, sold on the company’s website for US$ 99, incorporates OpenAI’s GPT-4o chatbot system.
• Print from the website FoloToy

Click the image to access.

• 30.12.2025
• DW
• In 2024, researchers at the University of California, Berkeley tested ChatGPT’s responses to different varieties of English dialects from countries such as India, Ireland, and Nigeria.
  
  The results show that the models tend to prioritize “standard” varieties of English (American or British). When prompted using dialectal forms, recurring issues emerged: stereotyping (19% more frequent), derogatory content (25% more frequent), lack of understanding (9% more), and condescending responses (15% more).
• Print from the news

Click the image to access.

• 16.01.2026
• CNN
• A woman reportedly lost €830,000 (more than R$5 million at current exchange rates) after falling victim to a scam in which she believed she was communicating online with actor Brad Pitt.
• Print from image attached to the news

Click the image to access.

• 31.08.2025
• G1
• A private school in the United States has adopted AI to personalize instruction, replacing teachers with digital programs capable of detecting each student’s level of knowledge and delivering accelerated learning, reducing formal class time to just two hours per day.
• Reproduction/Social media extracted from G1 website.

Click the image to access.

• 19.05.2025
• Época Negócios
• Atlassian’s CEO lays off 150 employees to replace them with AI.
• Print from the news.

Click the image to access.

• 07.07.2025
• MIT Technology Review Brasil
• The installation and rapid expansion of data centers in the state of Nevada, United States, has contributed to increasing water scarcity in the region. Driven by the growing demand for training and operating advanced artificial intelligence models, major technology companies have significantly expanded their computational infrastructure. These facilities require substantial amounts of water for continuous server cooling, intensifying pressure on already limited water resources in an arid environment.
• Image from the news website MIT Technology Review Brasil.

Click the image to access.

• 31.10.2024
• Deutsche Welle
• Scientists predict a thousandfold increase in electronic waste generated by artificial intelligence servers by 2030.
  
  Researchers warn that the rapid expansion of AI infrastructure is expected to dramatically increase electronic waste due to the accelerated replacement cycle of high-performance chips and specialized hardware. As AI systems demand ever more powerful processors, large volumes of outdated servers and components may be discarded, raising environmental concerns about toxic materials, recycling capacity, and long-term sustainability.
• Daniel Schäfer/dpa/picture alliance

Click the image to access.

• 06.05.2025
• ONU News
• High energy consumption by data centers.
  
  The growing deployment of artificial intelligence systems has significantly increased the energy demands of data centers worldwide. Training and operating advanced AI models require continuous, high-intensity computational power, leading to substantial electricity consumption and additional pressure on power grids. Experts caution that, without cleaner energy sources and efficiency improvements, AI-driven infrastructure could further intensify carbon emissions and environmental impacts.
• Unsplash/Taylor Vick

Click the image to access.

• 02.04.2025
• Desinformante
• The Studio Ghibli trend on ChatGPT 4.0 overlooked copyright concerns and highlights the need for media literacy.
• Print from the White House's oficial account on X

Click the image to access.

• 26.06.2025
• The Guardian
• Renowned authors such as Kai Bird and Jia Tolentino filed a lawsuit against Microsoft, accusing the company of training its AI model, Megatron, on nearly 200,000 pirated digital books - an allegedly unlawful practice that replicates the authors’ style, themes, and expressions.
• Craig T Fruchtman/Getty Images

Click the image to access.

• 26.08.2025
• Exame
• Japanese media outlets filed a lawsuit against AI search company Perplexity for copyright infringement. The Nikkei Group, which also owns the Financial Times, and Asahi Shimbun allege that Perplexity copied and stored articles from their websites without permission and ignored technical measures designed to prevent unauthorized accessJ
• Jaque Silva/Getty Images

Click the image to access.

• 05.02.2024
• Terra
• Cybercriminals used artificial intelligence technology to create an entirely fake videoconference meeting and deceive a finance department employee at a multinational company in Hong Kong. During the call, all other participants — including the supposed image and voice of the company’s chief financial officer, were AI-generated deepfakes, realistically replicating the appearance and speech of real individuals.
  
  Believing he was interacting with legitimate colleagues, the employee carried out approximately 15 bank transfers totaling around US$25–26 million (roughly R$125–130 million) to accounts controlled by the criminals. Hong Kong police later confirmed that the videoconference was completely fabricated.
• Print from the news.

Click the image to access.

• 07.02.2025
• Reuters
• In February 2025, Meta (Facebook) carried out a mass layoff of approximately 4,000 employees (around 5% of its global workforce) as part of a restructuring aimed at focusing on IA projects. Internal memos revealed that, at the same time, the company accelerated the hiring of machine learning engineers considered “critical to the business.”
• REUTERS/Yves Herman/File Photo. Image available on the news.

Click the image to access.

• 17.07.2025
• Reuters
• In June 2025, it emerged in the United States that the AI startup Anthropic had allegedly downloaded up to 7 million pirated books from the internet to train its Claude language model. A federal judge allowed a class-action lawsuit filed by authors against the company to proceed.
• REUTERS/Dado Ruvic/Illustration/File Photo. Image available on the news.

Click the image to access.

• 08.07.2024
• Mobile Time
• The intensive use of AI has increased the carbon footprint of big tech companies. In July 2024, Google’s sustainability report revealed that its greenhouse gas emissions grew by 48% over five years (reaching 14.3 million tons of CO₂ in 2023), partly due to the energy consumed by data centers to support AI resources.
• Print from the news.

Click the image to access.

• 22.05.2025
• Convergência digital
• In the United States, the Federal Reserve System (Fed) claims that artificial intelligence is driving job creation to nearly zero.
• Print from the news.

Click the image to access.

• 24.02.2025
• Exame
• AI-related pollution could cost up to US$ 20 billion to U.S. public health. A recent report indicates that low-income communities are the most affected, although the pollution spreads across regions and can impact a wide range of localities.
• Print from the news

Click the image to access.

• 13.08.2025
• Sic Notícias
• A study conducted in Poland and published this Tuesday indicates that the introduction of artificial intelligence (AI) appears to be making doctors less efficient at detecting colon cancer.
  
  According to the study, published in The Lancet Gastroenterology [&] Hepatology, regular use of AI seems to have “harmful effects on specialists’ skills,” making it one of the first studies to suggest a potential decline in medical competencies associated with AI reliance.
• Print from the news.

Click the image to access.

• 28.01.2026
• CNN
• Amazon announced the layoff of 16,000 employees amid the race to advance artificial intelligence. In a blog post published on Wednesday (28), the company stated that it needed to reduce bureaucracy in order to speed up its decision-making process.
• Print from the news

Click the image to access.

• 04.08.2025
• Aos Fatos
• Criminals steal images of sick children to carry out scams on Facebook.
• Image from the news portal "Aos Fatos".

Click the image to access.

• 14.08.2025
• CNN Brasil
• Meta’s chatbot behavior policies allowed the company’s artificial intelligence creations to engage in romantic or sensual conversations with children.
• Print from the news.

Click the image to access.

• 25.09.2024
• G1
• Students were expelled after using artificial intelligence to create fake nude images of teachers and classmates at a private school in Cuiabá.
• Image bank/Pixabay

Click the image to access.

• 20.03.2025
• Carta Capital
• Arve Holmen discovered that ChatGPT had falsely described him as having murdered two of his children. In response, he filed a formal complaint against OpenAI, alleging algorithmic defamation and a violation of privacy, as the system disseminated fictitious and highly offensive information about him.
• Print from the news.

Click the image to access.

• 27.07.2025
• Reuters
• Spanish police are investigating a 17-year-old accused of using artificial intelligence to create fake nude images of female classmates and distributing or selling them online. At least 16 students reported that manipulated images of them were circulating on social media.
  
  The case emerged after a student discovered photos and videos falsely depicting her “completely nude,” published through an account created in her name. According to Spain’s Civil Guard, the images were real photographs digitally altered to simulate nudity.
  
  The teenager is being investigated for corruption of minors. The case has intensified ongoing debate in Spain over regulating non-consensual AI-generated sexual content, including proposed laws criminalizing the creation and distribution of deepfake material. It also follows a similar 2023 incident involving minors.
• Print from the news.

Click the image to access.

• 05.03.2025
• Revista Oeste
• An international operation (codenamed Cumberland) exposed, in February 2025, a scheme involving the creation and sale of AI-generated child pornography.
• Print from the news.

Click the image to access.

• 02.07.2025
• The Spectador
• In July 2024, on the eve of the UK election, it was discovered that an anonymous website had published pornographic deepfake images featuring the faces of 30 female politicians from the United Kingdom.
• Picture by Jonathan Hordle Disp for ITV/via Getty Images

Click the image to access.

• 08.10.2025
• CNN Brasil
• Criminal groups in Rio de Janeiro have begun using AI-powered voice cloning in sophisticated phone scams. Posing as representatives of public agencies, they call victims and record short voice samples; they then use voice-cloning software to replicate the person’s voice and contact relatives requesting urgent money transfers, simulating kidnappings or emergency situations.
• Dragana_Gordic/Freepik

Click the image to access.

• 04.11.2025
• G1 Globo
• Criminals are expanding the use of artificial intelligence in fraud schemes. Elderly individuals with significant assets have become targets of scams involving cryptocurrencies.
• Print from the news.

Click the image to access.

• 03.11.2025
• G1 Globo
• The G1 news portal confirmed that a video showing a police officer spraying pepper spray in a woman’s face during an alleged protest in Brazil is false. The scene did not actually occur: the material was created using artificial intelligence and edited to appear as though it were part of news coverage of an event that never happened.
• Print from the news.

Click the image to access.

• 10.07.2025
• Portal Tela
• Grok, xAI’s chatbot, faced a significant crisis after generating violent and antisemitic posts. A change in its programming, intended to allow more “politically incorrect” responses, resulted in unacceptable content, including descriptions of violence. The incident led to the resignation of X’s CEO, Linda Yaccarino, after only two years in the role.
• Print from the news.

Click the image to access.

• 10.07.2025
• SIC Notícias
• The online security platform ESET warned about the circulation on TikTok of fake videos generated by artificial intelligence (AI) that imitate healthcare professionals and spread medical misinformation to deceive users of the application. This represents harm to the fundamental rights to health and information, as guaranteed by the 1988 Brazilian Federal Constitution.
• Print from the news.

Click the image to access.

• 22.12.2025
• Diego Almeida
• Fake academic articles created by artificial intelligence are already being cited in real, peer-reviewed scientific research. What began as a problem in student assignments has spread to academic journals and scientific databases.
  
  The danger is silent: AI-generated fabricated citations gain an appearance of legitimacy when repeated in new articles, confusing researchers, overburdening librarians, and putting the credibility of science at risk.
  
  Experts warn that this process of “citation laundering” may undermine trust in scientific research, compromise critical thinking, and contaminate the academic ecosystem if the use of AI is not accompanied by rigorous verification.
• Print from the post.

Click the image to access.

• 04.02.2026
• G1
• ChatGPT classifies Brazilians from the Southeast as “more intelligent” and portrays those from the North and Northeast as inferior, a study finds.
• Print from the article

Click the image to access.