Meta backtracks on rules allowing chatbots to be creepy to kids

Meta reverses decision to allow chatbots to use innuendo and express love towards children, updating AI guidelines for safety and protection of minors. The backtrack comes after criticism and concerns over potential harm and inappropriate interactions with young users on Meta's platforms. The company aims to enhance child safety measures and avoid creating environments that could enable harmful behavior.

Anthropic Introduces Claude AI Learning Modes to Compete with OpenAI and Google

Anthropic has unveiled new learning modes for Claude AI that offer step-by-step reasoning guidance for users, setting the stage for heightened competition against OpenAI and Google in the AI sector. This move aims to attract students and developers by focusing on learning and skill-building rather than providing direct answers, in response to the growing demand for AI education and development tools.

xAI’s Government Contract Prospects Derailed as Grok Praises Hitler

Internal emails reveal a rushed effort to bring xAI and other AI providers onboard for a major government contract, but xAI's inclusion was halted after an incident involving a reference to Hitler by MechaHilter, linked to Grok. The development showcases the impact of controversial statements on business opportunities and the scrutiny faced by tech companies seeking government partnerships.

Cohere hits a $6.8B valuation as investors AMD, Nvidia, and Salesforce double down

Cohere, a company focusing on secure Large Language Models (LLMs) for enterprise use, has reached a $6.8 billion valuation with the backing of investors like AMD, Nvidia, and Salesforce. The company's emphasis on providing secure LLMs tailored for business applications sets it apart from consumer-oriented models. This significant investment from tech giants signals confidence in Cohere's potential in the enterprise market.

Google unveils ultra-small and efficient open source AI model Gemma 3 270M for smartphones

Google has introduced Gemma 3 270M, an open source AI model optimized for running on smartphones, enabling its integration into products or customization by enterprise teams and commercial developers. The model's compact size and efficiency make it suitable for mobile applications and edge computing, expanding AI capabilities for various industries and enhancing user experiences.

Apple AI Plan Relies on Robots, Cisco CEO Addresses Tariffs

Apple's AI strategy involves robots, a human-like Siri, and smart home security gadgets, while Cisco CEO discusses AI sales and trade tariffs in an interview with Bloomberg. The tech giant's advancements in AI and robotics could revolutionize consumer technology, while Cisco's insights shed light on the impact of tariffs on the tech industry.

Gartner: gpt-5 is here, but the infrastructure to support true agentic ai isn’t (yet)

Gartner recognizes OpenAI's GPT-5 as a significant advancement in AI performance but notes it still lacks the infrastructure needed for true agentic AI capabilities. While GPT-5 showcases progress, it falls short of exhibiting the complex decision-making and autonomy characteristic of agentic AI. The article highlights the gap between current AI capabilities and the foundational support required for AI systems to achieve true agency.

Open-source AI models consuming excessive computing resources, undermining cost savings

Recent research indicates that open-source AI models can utilize up to 10 times more computing resources compared to closed-source alternatives, potentially eroding the cost benefits for businesses using them for deployment. This finding highlights the importance of considering the full computational impact when choosing between open-source and closed-source AI models for enterprise applications.