Apple introduces Foundation Models framework, empowering developers to harness on-device AI models created by Apple, enabling offline AI-driven experiences with features like guided generation and tool calling, natively supported by Swift and accessible with minimal coding.
Acknowledging its delayed AI advancements, Apple grants developers access to its on-device large language model, enabling AI integration in apps, yet still trails OpenAI and Google's AI capabilities despite OS updates leveraging its AI technology.
X, recently acquired by Elon Musk's AI company xAI, has updated its developer agreement to prohibit using X's content and API to train large language models, restricting competitors' access to its data for AI training purposes.
Cybersecurity researchers have discovered a new AI jailbreak technique called 'Bad Likert Judge' that can boost the success rates of malicious prompts bypassing large language models' safety guardrails by over 60%, enabling the generation of potentially harmful or illegal content.
Google Labs' Portraits experiment allows conversational AI interactions with trusted experts, starting with bestselling author Kim Scott's AI-powered coaching on communication and leadership.
Windsurf, a popular vibe-coding startup reportedly being acquired by OpenAI, says Anthropic significantly reduced its direct access to Claude AI models, forcing it to find alternative providers and potentially causing availability issues for users accessing these models on Windsurf's platform.
SAP announces massive Business AI expansion to over 400 scenarios by 2025, transforming Joule into an AI-powered interface with autonomous workflow agents that can execute complex tasks across multiple applications while eliminating manual prompt engineering.
Anthropic has launched Claude Explains, a blog featuring AI-generated content with human oversight, showcasing its Claude model's writing abilities while emphasizing collaboration between AI and human expertise to create educational resources across diverse topics.
Chinese tech giants and startups are gaining ground in the global race to develop AI agents that can automate complex tasks, with companies like Butterfly Effect, Alibaba, Zhipu, and ByteDance launching agents powered by their own AI models to compete with OpenAI and others.
Google quietly released an Android app called Google AI Edge Gallery that lets users download and run open-source AI models locally on their phones without an internet connection, enabling offline access to capabilities like image generation, question answering, and code writing.