Amazon research leads to ban on Anthropic's Fable model

Deloitte Japan is leveraging Cisco Foundation AI's open-source model to enhance its security operations. The model, with 8 billion parameters, analyzes alert logs and drafts summaries to help prioritize and handle security threats more efficiently.

In other news, concerns are rising about Chinese hackers increasing cyberattacks aimed at stealing advanced artificial intelligence technology in the United States. This growing threat is a focus for security firms.

The University of Utah is launching a new bachelor's degree in artificial intelligence, set to start in fall 2026. The program aims to meet the growing demand for AI professionals, building on the university's history with AI dating back to 1973.

Turkey has unveiled an AI action plan to support its digital transformation and technological capacity. The plan includes four main pillars: awareness, use, production, and governance, with a goal to train 5 million people in AI literacy within two years.

Humanoid robots are now capable of learning to differentiate themselves from others using sensory data. This self-other distinction enables better collaboration and task performance in human-robot environments.

The White House's AI policy has faced criticism for being disorganized and ineffective. Experts argue that the policy lacks coherence and clear goals, developed through various executive orders and directives.

Amazon's security research led to the White House's decision to ban Anthropic's Fable model due to security concerns. This ban has implications for the use of AI models by foreign nationals.

Google DeepMind's Logan Kilpatrick discusses the limitations of AI models, including the problem of models becoming overly specialized and failing to generalize. He explains how to design models and training methodologies that encourage exploration and flexibility.

Raquel Urtasun, CEO of Waabi, believes that Gen Z is better equipped to handle AI-first approaches than industry veterans. She prioritizes curiosity and adaptability in her hiring process, focusing on talent that can drive innovation in AI.

Key Takeaways

• Deloitte Japan uses Cisco Foundation AI's open-source model to enhance security operations.\n• Chinese hackers increase cyberattacks to steal AI technology in the US.\n• University of Utah introduces AI bachelor's degree starting in 2026.\n• Turkey launches AI action plan to train 5 million in AI literacy.\n• Humanoid robots learn self-other distinction for better collaboration.\n• US White House's AI policy criticized as disorganized.\n• Amazon research leads to ban on Anthropic's Fable model.\n• Google DeepMind's Kilpatrick discusses AI model limitations.\n• Waabi CEO Raquel Urtasun prioritizes Gen Z for AI talent.\n• AI models face challenges in generalization and specialization.

Deloitte Japan boosts security with Cisco AI model

Deloitte Japan is using Cisco Foundation AI's open-source model to improve its security operations. The model helps analyze alert logs and draft summaries, making it easier to prioritize and handle security threats. This approach streamlines tasks and improves detection quality. Deloitte Japan has developed three core workflows to support alert analysis, severity analysis, and suppression rule creation. The model has 8 billion parameters, making it relatively small and practical for deployment.

China boosts AI cyberattacks in the US

Chinese hackers are increasing cyberattacks aimed at stealing advanced artificial intelligence technology in the United States. This is a growing concern for security firms. The cyberattacks are focused on acquiring AI technology, which is a valuable asset for many countries.

University of Utah offers AI bachelor's degree

The University of Utah is introducing a new bachelor's degree in artificial intelligence, starting in fall 2026. The university has a strong history with AI, dating back to 1973. The new degree program aims to meet the growing demand for AI professionals. The university's Kahlert School of Computing will offer the program.

Turkey unveils AI action plan

Turkey has launched a new artificial intelligence action plan to support its digital transformation and technological capacity. The plan has four main pillars: awareness, use, production, and governance. The goal is to make Turkey a leading country in AI and digital production. The plan includes training 5 million people in AI literacy within two years.

Humanoid robots learn self-other distinction

Humanoid robots can now learn to differentiate themselves from others using sensory data. This self-other distinction capability enables better collaboration and task performance in human-robot environments. The robots use proprioceptive-visual correspondence to build predictive self-models.

US AI policy criticized as shambolic

The White House's AI policy has been criticized as disorganized and ineffective. The policy has been developed through various executive orders and directives, but experts argue that it lacks coherence and clear goals.

Amazon research leads to Anthropic ban

Amazon's security research reportedly led to the White House's decision to ban Anthropic's Fable model. The ban was due to security concerns, and it has implications for the use of AI models by foreign nationals.

DeepMind's Kilpatrick on AI model limitations

Waabi CEO bets on Gen Z for AI talent

Raquel Urtasun, CEO of Waabi, a self-driving truck unicorn, believes that Gen Z is better equipped to handle AI-first approaches than industry veterans. She prioritizes curiosity and adaptability in her hiring process.