OpenAI postpones Astra launch because it is "too dangerous" for cybersecurity
AI News
Искусственный интеллект

OpenAI postpones Astra launch because it is "too dangerous" for cybersecurity

SAPO Notícias2 September 2026 at 21:30

OpenAI decided to temporarily suspend the availability of Astra, its most advanced artificial intelligence model, due to serious digital security risks identified by the company.

The decision to postpone the launch is related to concerns about the technology's potential to be used maliciously in the cybersecurity domain. The company intends to ensure that the model does not perform harmful activities before being released to the public.

Astra represents one of OpenAI's most ambitious developments, but the security implications associated with a system of this magnitude prompted this preventive pause in its release.

This is not the first time OpenAI has adopted cautious measures regarding its products. The company has been implementing stricter security protocols as AI technology becomes more sophisticated and powerful.

Related articles

AI News

AI chatbots allow access to Russian disinformation

Several artificial intelligence chatbots, including ChatGPT and Claude, were found to reproduce and amplify information from Russian media outlets considered instruments of Kremlin disinformation. The study revealed that these AI systems can regurgitate pro-Russian narratives without adequately contextualizing them, representing a risk in the dissemination of disinformation during periods of geopolitical conflict.

Observador03/09/26, 00:55
AI News

Honey and the Algorithm

This opinion article warns of the dangers of early exposure of children to artificial intelligence and screen-centered childhood. The author argues that, although it is impossible to completely disconnect from AI, adults can and should limit children's access to large language models and advocate for traditional free play. According to the text, this type of play develops essential skills that no algorithm can teach, such as real friendships, empathy, frustration tolerance, and identity construction.

CNN Portugal02/09/26, 13:07
AI News

Artificial intelligence threatens to 'obliterate cultural difference': Pedro Abrunhosa

Portuguese musician Pedro Abrunhosa warned, in a speech at the "War and Peace" event, that artificial intelligence threatens to "obliterate cultural difference" between peoples and societies. According to Abrunhosa, the true creative process is born from the unconscious, from the inexplicable and from profound human experiences — elements that technology, no matter how advanced, cannot replicate.

Now02/09/26, 12:31
Madeira ganha espaço na Inteligência Artificial
AI News

Madeira Gains Ground in Artificial Intelligence

The island of Madeira hosts an international conference bringing together about 120 specialists from 28 countries in Funchal, marking the growing relevance of the region in the field of Artificial Intelligence and positioning the archipelago as an emerging hub in the technology sector.

SAPO02/09/26, 11:24