More results...
The amount of data being shared by businesses with GenAI apps has exploded, increasing 30x in one year, according to Netskope. The average organization now shares more than 7.7GB of data with AI tools per month, a massive jump from just 250MB a year ag…
Measure the different level of risk inherent to all gen-AI foundational models and use that to fine-tune the operation of in-house AI deployments.
The post New AI Security Tool Helps Organizations Set Trust Zones for Gen-AI Models appeared first on Sec…
Two Microsoft researchers have devised a new jailbreak method that bypasses the safety mechanisms of most AI systems.
The post New CCA Jailbreak Method Works Against Most AI Models appeared first on SecurityWeek.
As AI technology advances, cybercriminals create more personalized and convincing scams. This includes mimicking voices, deepfake videos, and highly convincing phishing emails that are difficult to spot. Phishing, deepfakes, and voice cloning are among…
AIceberg has launched a solution that helps governments and enterprises with the safe, secure and compliant adoption of AI.
The post AIceberg Gets $10 Million in Seed Funding for AI Security Platform appeared first on SecurityWeek.
Organizations have zero visibility into 89% of AI usage, despite security policies according to a LayerX report. 71% of connections to GenAI tools are done using personal non-corporate accounts. Among logins using corporate accounts, 58% of connections…
In a lawsuit targeting cybercriminals who abuse AI services, Microsoft has named individuals from Iran, the UK, China and Vietnam.
The post Microsoft Names Suspects in Lawsuit Against AI Hackers appeared first on SecurityWeek.
GPT-4.5는 주로 비지도 학습 확장에 중점을 둔 모델로, ‘생각하기 전에 응답하지 않는다’라는 가치를 추구한다. 이 모델은 오픈AI가 이전에 공개했던 o1, o3-mini와 같은 추론 중심 모델과는 다르다. AI 개발 및 훈련 영역에서 비지도 학습은 넓은 지식과 패턴 인식, ‘직관’에 가까운 능력을 강화하는 반면, 추론은 논리적 사고, 문제 해결, 단계적 분석 능력을 강화하는 특징이 있다.
오픈AI…
‘생성형 AI의 새로운 지평을 넘어서’를 주제로 열린 이번 행사에서 알테어는 기업이 sLLM에 주목해야 할 이유를 강조했다. 알테어의 이상헌 이사는 “대형 언어 모델(LLM)이 빠르게 발전하고 있지만, 기업 환경에서는 데이터 보안, 비용, 할루시네이션(환각) 문제 해결이 여전히 중요하다”라며 “특히 기업이 자체 데이터를 안전하게 활용하고 맞춤형 AI 솔루션을 구축하기 위해서는 보다 경량화된 모델이 필요하다”…
Trends in cybersecurity across 2024 showed less malware and phishing, though more social engineering. CrowdStrike offers tips on securing your business.