Artificial Intelligence News Live: Users can disable integrated Gemini AI features in Google services like Gmail and Docs

Google provides an option for users to disable its integrated Gemini AI features within services such as Gmail, Docs, and Photos, according to a report in ZDNet. While the AI functionalities are not always active by default, the prompts and availability of Gemini tools can be turned off through a settings option. The report notes that while the process to opt-out is straightforward, the setting itself may not be easily discoverable for all users, reflecting a growing demand for user control over pervasive AI integrations.

Google has enabled direct integration of its Gemini AI within the Chrome browser, allowing users to access the model from any webpage, according to a report in ZDNet. By typing "@gemini" followed by a prompt in the address bar, users can leverage the AI to perform tasks that are contextually aware of the current page's content, such as summarizing articles, explaining concepts, or drafting text. This development is part of a broader strategy to embed AI across Google's ecosystem and intensifies the competition with other browsers like Microsoft Edge, which has a similar Copilot integration. While the feature aims to enhance productivity, it also involves sending browsing context to Google's servers, and users should be aware of the AI's limitations, including the potential for generating incorrect information.

Indian startup QpiAI has secured $32 million in a new funding round co-led by the Indian government, according to a report in TechCrunch. The company integrates artificial intelligence with quantum computing to develop solutions for enterprise applications. This investment aligns with India's strategic goal to advance its national capabilities in quantum technologies, which is supported by the country's National Quantum Mission established in 2023.

A group of former top researchers from Google has developed a new type of artificial intelligence agent, according to a report in Wired. The new venture's stated mission is to advance AI by teaching models to better understand and build software code. This focus on code generation is presented as a pathway toward the long-term goal of developing superintelligent AI.

Research from Amazon indicates that artificial intelligence is already augmenting entry-level technology jobs, according to a report in ZDNet. The findings suggest AI is being integrated into workflows to assist and enhance human efficiency rather than for outright job replacement. The research also highlights an increasing rate of growth in the number of people qualified for these augmented roles, pointing to a rising demand for workers who can effectively utilize AI tools.

Google is rolling out a new feature for its Discover feed that will use artificial intelligence to provide summaries of news stories, as per a report in ZDNet. This development is designed to allow users to quickly grasp the main points of an article directly within their personalized feed without needing to click through to the full page. The feature is part of Google's broader strategy to integrate generative AI across its products, similar to the AI-powered overviews seen in its Search Generative Experience (SGE).

Google has rolled out a new AI-powered feature designed to call local businesses on behalf of a user to inquire about information such as availability and pricing, according to a report in TechCrunch. This function represents an evolution of the company's agentic AI capabilities, which aim to perform multi-step tasks autonomously. In a parallel development, Google has also integrated its advanced Gemini 2.5 Pro large language model into its 'AI Mode' to enhance the platform's capabilities.

AI data labeling firm Scale AI is laying off 200 employees and 500 global contractors as part of a major restructuring, The Verge said in a report. In an internal email, CEO Jason Droege stated the company had "ramped up our GenAI capacity too quickly," which created inefficiencies and redundancies. The company will now streamline its generative AI business from 16 units down to five and reorganize its go-to-market team. The changes come one month after Meta invested $14.3 billion for a 49% stake in the company and hired former CEO Alexandr Wang to lead its new superintelligence lab. A company spokesperson confirmed that Scale AI plans to increase investment and hire in other areas, such as its Enterprise and Public Sector divisions, in the second half of 2025.

A new application for macOS, Substage, allows users to manage files and folders using natural language commands, according to a report in Wired. The tool integrates with the Finder and uses artificial intelligence to interpret plain English requests, translating them into executable commands. Users can type instructions to perform tasks such as converting file types, creating zip archives, and organizing folders, simplifying complex operations that would otherwise require scripting or manual steps.

China is accelerating its push to become a leader in artificial intelligence through a state-led industrial policy involving massive investment estimated in the hundreds of billions of dollars, The New York Times says in a report. This long-standing strategic approach aims to close the technology gap with the United States, an effort intensified by U.S. export controls on advanced semiconductors. While China has made significant progress in AI applications and developing its own large language models, the U.S. is generally considered to maintain an advantage in foundational AI research and cutting-edge chip technology. The strategic competition between the two nations is driving a global push for technological self-sufficiency and reshaping technology supply chains.

Chinese autonomous driving firms, including Baidu, Pony.ai, and WeRide, are rapidly scaling commercial robotaxi services and expanding internationally, challenging US pioneers like Waymo and Tesla, according to a report in IEEE Spectrum. Baidu's Apollo Go service now operates 1,000 vehicles across 15 cities, delivering 1.4 million rides in the first quarter of 2025. These companies leverage significant cost advantages, with Baidu producing vehicles for as low as $28,000, and benefit from training their AI in China's complex urban traffic. All three are now expanding into the Middle East and Europe. While Waymo remains a leader in the US, Tesla's recent pilot program is considered far behind in large-scale commercial deployment. Key challenges for the Chinese firms include a fragmented regulatory landscape, a strategic reliance on US-made Nvidia chips, and achieving profitability.

High-profile researcher Jason Wei, who worked on OpenAI's 'o3' and deep research models, is leaving the company to join Meta's superintelligence lab, Wired says in a report. A source also confirmed that Wei's colleague, Hyung Won Chung, is making the same move. The departures highlight the ongoing competition for top talent among major technology firms as they pursue advancements in artificial intelligence.

A new study from Google's DeepMind reveals that large language models (LLMs) can abandon correct answers when put under pressure during multi-turn interactions, a finding that poses challenges for the reliability of complex AI systems, according to a report in VentureBeat. The research identified a "confidence paradox" where models can be simultaneously stubborn, showing high confidence in wrong answers, while also being easily swayed from correct ones when prompted iteratively or presented with conflicting information. This behavior highlights a key vulnerability for developing dependable AI applications, especially those requiring multi-step reasoning or sustained conversation.

Microsoft is rolling out an update for its Copilot Vision AI to Windows Insiders that expands its capabilities to scan a user's entire screen, a specific browser, or an application window, according to a report in The Verge. This feature must be manually activated by the user, similar to screen sharing in a video call, and is distinct from the 'Recall' feature as it does not automatically take snapshots of screen activity. Once activated, Copilot can analyze on-screen content to provide insights, answer questions, or offer guidance on various tasks such as creative projects or navigating a game.

Germany plans to significantly ramp up its use of artificial intelligence by the end of the decade to support its economy and enhance its global competitiveness in key technologies, as per a report in Reuters. The plan, detailed in a government strategy document, represents an "AI offensive" designed to accelerate and expand upon the country's national AI strategy first established in 2018. The initiative aims to close perceived gaps with leading nations in areas such as venture capital investment, AI patents, and the broad adoption of AI across industries, with a particular focus on integrating AI into Germany's strong manufacturing and engineering sectors. This strategic push will be implemented within the regulatory framework of the European Union's recently finalized AI Act.

Oracle will invest $3 billion in Germany and the Netherlands over the next five years to expand its infrastructure for artificial intelligence and cloud services in the European market, Reuters said in a report. The investment is aimed at strengthening Oracle's competitive position and meeting the growing demand for high-performance computing and localized data services, driven by the rapid adoption of AI and European data sovereignty regulations.

U.S. startup funding increased by 75.6% in the first half of 2025, fueled by sustained investment in the artificial intelligence sector, according to a report in Reuters. The data, from a PitchBook report, indicates the funding pace is on track for the second-best year on record for U.S. startups. This surge in AI-focused investment contrasts with the broader market, where many venture capital firms have reportedly faced difficulties in raising money for their own funds, highlighting a bifurcated environment where AI is the primary driver of investment.

BMW Group announced a partnership with Chinese autonomous driving company Momenta to co-develop advanced driving assistance systems tailored for the Chinese market, Reuters said in a report. The collaboration is aimed at creating technology specifically optimized for China's unique road conditions and traffic patterns, while also ensuring compliance with the country's strict data security regulations.

Artificial intelligence research company Moonvalley has raised an additional $84 million in a funding round led by existing investor General Catalyst, Reuters said in a report. The new funding was announced on Monday, approximately one week after the company released its first video generation AI model. The capital is intended to bolster the development of its AI video tools.