AI Innovations and Legal Battles: A Month in Review
This month has been a whirlwind of advancements and legal drama in the AI industry.
Alibaba’s New AI Models
Alibaba kicked off the month by unveiling Qwen3, a new family of AI models that reportedly match, and sometimes surpass, the performance of leading models from OpenAI and Google. This marks a significant step forward in the AI race, showcasing Alibaba’s commitment to pushing the boundaries of technology.
Meta’s AI Assistant App
Meta has launched a dedicated application for its AI assistant, following its integration into WhatsApp, Instagram, Facebook, and Messenger. What sets Meta’s version apart is its ability to utilize extensive user data from Facebook and Instagram, personalizing experiences based on individual profiles and interaction history. However, this innovative app is currently available only in the United States and Canada, leaving other regions, including New Zealand, waiting for its release.
Elon Musk vs. OpenAI: The Legal Saga Continues
The legal dispute between Elon Musk and OpenAI, along with its CEO Sam Altman, took a significant turn when a US judge denied OpenAI’s request to dismiss the case. Musk, who played a crucial role in founding and financing OpenAI, claims that the company has strayed from its original non-profit mission by transitioning to a for-profit structure, which he argues constitutes fraud and breach of contract. Additionally, Musk has implicated Microsoft due to its substantial support of OpenAI.
OpenAI’s Shift to Public Benefit Corporation
In response to the ongoing legal battle, OpenAI announced that it had abandoned its plan to become a for-profit company after discussions with civic leaders and the attorneys general of California and Delaware. OpenAI’s commercial arm will now become a Public Benefit Corporation (PBC), joining others like Anthropic and xAI. This move aims to align OpenAI’s operations with its original mission and address concerns raised by Musk and other stakeholders.
Netflix’s AI-Driven Search Tool
Netflix is experimenting with a new AI-driven search tool that allows users to describe their preferences in simple terms, such as “comforting” or “slightly dark but not overly so, perhaps with dry humour”. This tool, powered by OpenAI’s ChatGPT technology, aims to simplify the process of selecting content to watch, enhancing user experience and engagement.
Google’s AI Futures Fund
Google has launched the AI Futures Fund to support startups using its DeepMind tools. The fund offers early access to AI models, Google Cloud credits, and guidance from DeepMind and Google Labs. Some startups may even receive direct investment, fostering innovation and growth in the AI sector.
Google’s AI Mode in Search Engine
In addition to the AI Futures Fund, Google has introduced AI Mode in its search engine. This chatbot-style tool integrated into the Search function allows users to pose questions, ask follow-up queries, and receive comprehensive responses directly from its AI system, making information retrieval more intuitive and efficient.
Nvidia’s NVLink Fusion
Nvidia CEO Jensen Huang announced NVLink Fusion at Computex 2025, enabling other companies to use non-Nvidia chips alongside Nvidia’s through NVLink. This allows firms like Fujitsu and Qualcomm to combine Nvidia GPUs with their CPUs in AI data centres. MediaTek, Marvell, and Synopsys are already partners, while AMD, Intel, and Broadcom remain uncommitted. Analysts believe it’s a smart move, strengthening Nvidia’s GPUs in hybrid AI systems despite a potential decrease in CPU market share.
OpenAI’s Acquisition of io
OpenAI has acquired io, a hardware startup founded by Jony Ive and ex-Apple engineers, for nearly $6.5 billion. Ive’s design firm, LoveFrom, will manage all OpenAI product designs. Additionally, around 55 io employees, including engineers and manufacturing specialists, are joining OpenAI, bolstering its hardware capabilities.
Anthropic’s Voice Mode for Claude Chatbot
Anthropic is launching a voice mode for its Claude chatbot app, enabling full spoken conversations in English using the Claude Sonnet 4 model. This feature, arriving in beta, offers five voice options, supports text-to-voice switching, and provides key point summaries along with transcripts during chats, enhancing user interaction and accessibility.