Latest news and stories about ai regulation in technology in Portugal for expats and residents.
Around 50 members of the European Parliament, including 46 women, have been targeted by non-consensual AI-generated deepfake pornography according to a new report, highlighting growing concerns over synthetic explicit content impacting approximately one in six female EU lawmakers.
Professor José Moreira of Portucalense University discusses the legal and ethical implications of artificial intelligence in hiring and firing processes.
Meta will postpone the implementation of its new commercial terms from October 30 to November 20 to address concerns from digital publishers and advertisers, reports Reuters. The measure was being contested for opening the door to web scraping, including of news sites, and their potential use for...
Regulating AI is necessary. But an educational policy is not measured by the number of recommendations published nor by the speed with which it adopts the language of innovation.
The European Union Agency for Cybersecurity (ENISA) is testing artificial intelligence (AI) models from the French firm Mistral to evaluate their hacking capabilities, at a time when European institutions are looking to understand how AI can be used both to launch cyberattacks and to prevent them. The agency is particularly interested in...
Brad Smith describes AI as “a tool and a weapon” that requires regulation to preserve human control.
The unauthorised use of copyrighted books to train AI models is now a growing concern and reality within the Portuguese literary and technological landscape.
The discussion on AI should not be between innovation and security: a competitive Europe needs both
One in six female MEPs has been the target of AI-generated deepfake pornography, a study reveals. Platforms have until December to comply with the ban on systems that generate these images.
Ethical reasons may be being used not for their relevance, but as a strategy for a pause that is more 'technical' than 'ethical', dictated by competitiveness rather than prudence. Opinion by Maria do Céu Patrão Neves
The creation of a global oversight body and the need to keep Artificial Intelligence (AI) 'under human direction, supervision and control' are two of the directives contained in a document published by the Politico website this Monday, which has so far been signed by 22 countries. However, one of the ideas that has been defended in various forums—the slowing down of AI development—is absent from this letter.
The debate gained new momentum after Anthropic CEO Dario Amodei publicly advocated for slowing down the pace of AI development.
During the State of the Union address, the European Union leader pointed out several areas where Artificial Intelligence can be applied, including factories, hospitals, precision agriculture, energy grids, autonomous driving, and defence.
An analysis of the concerns, proposed regulations, and pledges regarding the pace of artificial intelligence development.
The advancement of autonomous AI agent systems—capable of using tools, executing code, and navigating the web independently—has introduced a new paradigm in cybersecurity. Recent incidents involving leading labs, such as OpenAI and Anthropic, demonstrate that the security risk is no longer theoretical: agents with offensive capabilities can bypass sandboxes and directly impact real-world infrastructure. These events, including unauthorized communication between agents and unauthorized access to third-party systems like Hugging Face, highlight the urgent need for stricter containment, robust governance, and the development of automated defensive AI to counter the growing threat of autonomous cyberattacks.
Several incidents involving AI systems have exposed new capabilities and security flaws. Now, the leaders of the companies developing them are arguing, for the first time, that it is necessary to slow down.
Regarding the joint request by Dario Amodei, Sam Altman, and Elon Musk for a pause in AI development until better safety measures are in place, Portas believes that Donald Trump's statements on the subject have “dynamited” any possibility of regulation imposed by the American government.
Dario Amodei, CEO of Anthropic, presents a series of proposals for “frontier” companies like his own. Because AI as it stands, he says, poses serious risks. Read the full article.
Months after its launch in April, Anthropic has granted access to Mythos, its artificial intelligence (AI) model specialised in cybersecurity, to the European Union Agency for Cybersecurity (ENISA), a European Commission spokesperson confirmed to Bloomberg. In the United Kingdom, the AI Safety Institute (AISI), the British institute responsible for testing these...
An analysis of the integration of Artificial Intelligence in the judicial system, emphasizing that while AI can assist with administrative and research tasks, the core responsibility of judging must remain a human function. The text highlights the constitutional necessity of judicial independence, the importance of transparency regarding AI usage in legal proceedings, and the need for judges to retain the authority to override algorithmic suggestions to ensure justice and human dignity.
Volker Turk criticises the delay in creating rules for the technology
The European Parliament is holding a public hearing in Brussels this Wednesday, September 2, regarding the impact of artificial intelligence (AI) on cultural and creative sectors. Led by Portuguese MEP Hélder Sousa Silva, the initiative aims to gather input from over thirty organizations across the European cultural and digital ecosystem to inform future regulations that protect cultural activities. The hearing covers four thematic blocks: music, audiovisual, publishing and media, and digital platforms, with a focus on transparency, content labeling, and ensuring innovation does not compromise the rights of creators.
The European Commission has opened a consultation on guidelines that will help companies and public entities classify high-risk artificial intelligence systems. The process includes uses related to employment and recruitment, at a time when selection and assessment tools are beginning to be used by more organisations.
Andrew Bailey warns that frontier models alter the “speed, scale and dimension” of cyber risk and calls for urgent measures to ensure the “safe and responsible” use of artificial intelligence.
Bernardo Caldas, an expert in Artificial Intelligence, analyses the agreement Meta signed following accusations from 29 US states, which forces the company to pay 14 billion euros and make changes to its application rules regarding young people.
Conspiracy theories, climate change denial, and the assistance of Artificial Intelligence defined disinformation and fake news throughout June and July. The monthly report from the European Digital Media Observatory (EDMO) reveals that the percentage of fake news created or manipulated with AI increased again during these months, reversing the decline observed in April and May.
In public offices, employees are increasingly using AI tools on personal devices to draft official responses without clear guidelines or oversight. Despite the existence of the EU AI Act, national agendas, and internal tools like AMALIA, the implementation remains ad hoc and unregulated. The author argues that the State is failing to provide necessary training and formal policies, leading to a lack of accountability and a potential erosion of public trust in administrative processes.
Artificial intelligence allows for a person to be stripped or depicted in sexual acts. But should such a vile violation of sexual self-determination be limited to administrative offences, or is a criminal law response justified?
Anthropic will start including invisible watermarks in texts generated by new Claude models. The measure is part of the commitments made by the company under the European Union's new transparency rules.
