Sitemize üye olarak beğendiğiniz içerikleri favorilerinize ekleyebilir, kendi ürettiğiniz ya da internet üzerinde beğendiğiniz içerikleri sitemizin ziyaretçilerine içerik gönder seçeneği ile sunabilirsiniz.
Sitemize üye olarak beğendiğiniz içerikleri favorilerinize ekleyebilir, kendi ürettiğiniz ya da internet üzerinde beğendiğiniz içerikleri sitemizin ziyaretçilerine içerik gönder seçeneği ile sunabilirsiniz.
Üyelerimize Özel Tüm Opsiyonlardan Kayıt Olarak Faydalanabilirsiniz
OpenAI announced that it has deliberately slowed the development of its upcoming artificial‑intelligence system, Astra, citing serious security concerns. According to the company, the model has crossed what it calls a “critical cybersecurity threshold,” meaning it could independently discover and launch cyber‑attacks against real‑world systems that are typically well defended.
Astra is positioned as the next generation of large language models, designed to handle highly complex tasks and make decisions that resemble human reasoning. During internal testing, researchers observed that the model was not only capable of analyzing sophisticated attack scenarios but also of generating actionable exploit strategies without human prompting.
OpenAI describes the critical threshold as the point at which an AI can autonomously produce attack vectors that even seasoned security professionals find difficult to detect or mitigate. In other words, the model moves from being a theoretical risk to a practical one that could be weaponized.
In response, the organization is applying its “responsible AI” framework: adding extra layers of oversight, imposing usage limits, and integrating human‑in‑the‑loop controls before any further scaling. OpenAI stresses that it does not intend to abandon Astra’s potential benefits, but rather to reshape its development path so that safety measures keep pace with capability.
Industry analysts see the move as a wake‑up call for the broader AI community. Some call for stricter regulation and greater transparency around models that can be repurposed for malicious ends, while others warn that excessive caution could slow innovation and delay useful applications.
OpenAI plans to work closely with academic partners, security experts, and policy makers to draft standards that curb misuse while preserving the technology’s promise. The company promises a detailed update on new safety protocols and a revised roadmap for Astra in the coming months.
Source: TechCrunch
ETİKETLER
BENZER İÇERİKLER
OKUYUCU YORUMLARI
OpenAI Halts Astra Model Progress Over Security Fears
Yorum Yaz