Tag: AI security

Unveiling a concerning vulnerability in AI systems, researchers discover the potential for dangerous content generation. Urgency to prioritize AI security and ethical development grows. #AIsecurity

ChatGPT's AI tool poses security risks to organizations as employees unknowingly share confidential information. Banyan Security's SWG solution can categorize data and block access to generative AI sites, ensuring network security.

Google introduces Secure AI Framework to protect businesses' AI models from unauthorized access. Microsoft has also committed to AI transparency and ethics by unveiling three new AI Customer Commitments. Read more here.

Google emphasizes the need to prioritize security measures for Artificial Intelligence against potential misuse. A six-pillared approach includes expanding threat intelligence work, employing automation in response processes, and regular security measures reviews. Companies should work with experts in risk management to ensure a holistic approach to protect all stakeholders. Google is working with governments and customers to address these concerns.

At the 11th International Conference on Learning Representations (ICLR), teams from around the world presented their research on efficient and effective neural network trojan removal techniques. The winning team from The Harbin Institute of Technology in Shenzhen won the $5,000 USD first-place prize with a 98.14% poisoned accuracy rate and a 0.12% attack success rate. The IEEE Computer Society created the Emerging Technology Fund to explore AI and machine learning security, and make these technologies more available and dependable for everyone.