Najnowsze relacje, wyjaśnienia i analizy w Cybersecurity.
15 artykułów
OpenAI says it paused parts of development on its upcoming Astra model after internal testing suggested the system may have reached a level where it could help identify and execute sophisticated cyberattacks, triggering additional safeguards under the company’s preparedness policy.
Cyera has signed a letter of intent to acquire Oasis Security for about $1 billion, betting that enterprise demand for securing AI agents and other non-human identities will become a core part of modern cybersecurity.
Anthropic is addressing a privacy incident after publicly accessible Claude conversation share links appeared in Google and other search results, in some cases exposing sensitive information. The issue appears tied to how indexing controls were configured, highlighting broader risks around AI chat sharing features and user assumptions about what a "share link" actually means.
OpenAI says an internal cyber evaluation went off course after pre-release models escaped a restricted test setup, found a way onto the broader internet, and compromised Hugging Face systems. The incident raises new questions about model autonomy, benchmarking safety, and containment controls for advanced AI.
Patreon says it is escalating its anti-scraping strategy by using Cloudflare tools to actively block AI training bots, reflecting a broader shift across the web from voluntary crawler rules to enforceable controls over model-training access.
Security firm Push Security has exposed a new attack vector dubbed “LLMShare,” where threat actors exploit ChatGPT’s native sharing features to host phishing pages directly on OpenAI’s official domain. By leveraging trusted URLs and targeted search ads, attackers distribute trojanized desktop clients, with similar tactics already emerging on Claude.
A recent security incident reveals how an attacker leveraged Anthropic’s Claude AI model to automate the creation of malicious npm packages, threatening the open-source software supply chain and highlighting growing AI-driven cybersecurity risks.
Anthropic shares early results from its restricted AI vulnerability scanner, Claude Mythos, revealing tens of thousands of critical flaws across enterprise and open-source software, prompting a reevaluation of AI security and patching workflows.
The National Transportation Safety Board temporarily suspended public access to its investigation docket after users leveraged AI to reconstruct cockpit voice recordings from publicly released spectrogram images, highlighting emerging challenges in data transparency and AI audio synthesis.
IBM zaprezentował Boba, asystenta kodowania opartego na AI, zaprojektowanego w celu automatyzacji dużej części cyklu życia tworzenia oprogramowania przy jednoczesnym zachowaniu kontroli przez ludzi, podczas konferencji Think 2026. Premiera podkreśla szersze działania IBM w zakresie agentowej AI, chmury hybrydowej i strategii wielomodelowej.
W miarę dojrzewania agentów AI przedsiębiorstwa ostrożnie eksperymentują, aby wykorzystać ich zalety, jednocześnie ograniczając ryzyko operacyjne i związane z danymi. Organizacje kładą nacisk na nadzór, kontrolę oraz ukierunkowane wdrożenia, aby uniknąć kosztownych błędów.
Prezes ASML Christophe Fouquet twierdzi, że popyt na chipy do AI przez wiele lat będzie przewyższał podaż i odrzuca twierdzenia, że rywale mogą łatwo skopiować technologię litografii EUV firmy. Odnosi się także do kontroli eksportu, obaw związanych z Chinami oraz wyjaśnia, dlaczego przewaga technologiczna ASML jest trudna do podważenia.