Artificial Intelligence News Live: Microsoft to cut up to 9,000 jobs while increasing investment in AI

Microsoft plans to reduce its global workforce by up to 9,000 positions, or 4% of its staff, as part of a strategic shift to increase investment in artificial intelligence, BBC says in a report. The company attributed the decision to a need to align its cost structure with revenue amid current macroeconomic conditions, while also reallocating resources and talent toward high-growth areas, specifically citing AI.

OpenAI has publicly stated that it does not endorse and was not involved in the sale of 'OpenAI tokens' on Robinhood, according to a report in TechCrunch. In a post from its official newsroom account, the artificial intelligence company clarified that these tokens will not provide consumers with equity or stock in OpenAI. The company emphasized its disavowal of Robinhood's effort and confirmed it had no role in facilitating the token sale.

OpenAI has signed an estimated $30 billion data center deal with Oracle to secure the massive computing power required for developing more powerful artificial intelligence models, according to a report in the Financial Times. The agreement is reportedly part of the "Stargate" joint venture, a project involving OpenAI and SoftBank aimed at building a network of AI supercomputers. Under the deal, Oracle will provide its Oracle Cloud Infrastructure (OCI) services to host and operate a significant portion of the infrastructure needed to meet the escalating demand for high-performance computing in the AI industry.

Hackers are leveraging generative AI to rapidly create highly convincing phishing websites that mimic login pages for major enterprise platforms like Okta and Microsoft 365, ZDNet says in a report. Research from Okta found that a functional phishing site could be generated in as little as 30 seconds, significantly lowering the barrier for cybercriminals to launch sophisticated social engineering attacks. These platforms are targeted because a single compromised account can grant attackers broad access to a company's network and sensitive data. The development highlights the dual-use nature of AI, prompting calls for stronger security measures like multi-factor authentication and advanced, AI-driven threat detection systems to counter these evolving threats.

Materials experts have used artificial intelligence to formulate new paint coatings that can keep buildings between 5C and 20C cooler than conventional paint after sun exposure, according to a report in The Guardian. The development, an example of machine learning accelerating materials science, could help reduce the urban heat island effect in cities and lower energy use from air conditioning. Researchers state the new coatings could also be applied to vehicles, trains, and electrical equipment to improve cooling as global temperatures rise.

Referral traffic from AI chatbots like ChatGPT to news publishers is on the rise, but it is not significant enough to counteract the much larger decline in traffic from traditional search engines, according to a report in TechCrunch. This trend coincides with a sharp drop in overall organic traffic to news sites, which has fallen from a peak of over 2.3 billion visits in mid-2024 to under 1.7 billion currently. The development highlights the challenge for publishers as users increasingly get information directly from AI platforms, leading many news organizations to explore content licensing deals with AI companies as an alternative revenue stream.

Streaming service Crunchyroll reportedly used unedited AI-generated subtitles for its new anime series, 'Necronomico and the Cosmic Horror Show', The Verge says in a report. Viewers discovered the subtitles were filled with significant typos, grammatical mistakes, and explicit mentions of ChatGPT, indicating they were likely produced by an AI model without human review or editing. The errors, such as the phrase 'Is gameorver. if you fall, you …', have led to criticism over the quality and implementation of AI in content localization.

The International Association of Insurance Supervisors (IAIS) has published an Application Paper on the supervision of artificial intelligence in the insurance sector on July 2, 2025, as per a report in International Association of Insurance Supervisors. The document provides non-binding guidance and best practices for insurance supervisors globally. It aims to help regulators apply supervisory principles to the growing use of AI in areas like claims processing, risk assessment, and pricing, while addressing potential risks such as algorithmic bias and ensuring fair outcomes for consumers.

In an effort to better understand human cognition, scientists have developed an artificial intelligence model that replicates human thought processes, including inherent flaws and biases, The New York Times says in a report. The researchers trained a large language model (LLM) using a specialized dataset of 10 million questions from psychology experiments. The primary goal of this research is to use the AI as a computational model for psychological and cognitive science, allowing for new ways to test hypotheses about how the human mind works, rather than to develop a commercial product.

Energy company Chevron is utilizing agentic artificial intelligence systems to significantly reduce the time for a specific data-intensive task from 30 days to one, according to a report in VentureBeat. The company's AI architecture is designed to process petabytes of operational data using a hybrid model that combines scalable cloud computing for large-scale analysis with edge computing for low-latency processing near the data's source. These agentic systems, which can act with a degree of autonomy, are being deployed to make decisions and optimize critical operations in the energy sector.

Baidu is overhauling its primary search engine in China, introducing new artificial intelligence features and a voice function in what is described as its most sweeping revamp in years, according to a report in Bloomberg. The move involves integrating advanced AI capabilities, likely powered by its Ernie Bot large language model, to enhance the user experience with more conversational interactions and direct answers. This update aligns Baidu with the global trend of major search engines embedding generative AI into their core services, intensifying the technological competition within China's digital market.

A new AI travel agent, developed by a startup founded by former engineers from Google, Meta, and Waymo, can generate complete trip itineraries for users at no cost, as per a report in ZDNet. The service uses generative AI to plan activities, accommodations, and transport routes, although it does not handle direct bookings. While the tool is currently free, the company's long-term monetization strategy has not been specified. This development is part of a broader trend of advanced AI integration in the travel industry and a talent shift from major technology companies to specialized AI startups.

SoftBank Group's planned acquisition of semiconductor designer Ampere Computing is now facing a detailed US government probe, potentially jeopardizing the company's artificial intelligence strategy, according to a report in Bloomberg. The investigation, likely by the Committee on Foreign Investment in the United States (CFIUS), centers on national security concerns related to Ampere's role in designing energy-efficient server chips. These processors are critical for the data centers that power cloud computing and artificial intelligence workloads, and the acquisition is a key part of SoftBank's strategic pivot towards AI.

The CEO of data-labeling startup Scale AI has stated the company will maintain its independence and focus on expansion despite Meta Platforms recently acquiring a 49% stake, according to a report in Bloomberg. The statement addresses the significant investment from the social media giant, which highlights the strategic importance of high-quality data for training advanced artificial intelligence models. Scale AI provides essential data annotation services, a foundational component for developing AI systems like those Meta is building.

Microsoft has secured a cloud computing agreement to provide its artificial intelligence technology to the Premier League, according to a report in Bloomberg. The pact will involve the English football league utilizing Microsoft's Azure cloud infrastructure and AI services. This partnership is designed to enhance areas such as fan engagement and data analytics, while providing Microsoft a significant platform to showcase its AI capabilities to the league's large global audience.

The operator of the AI-powered "nudify" app Clothoff is reportedly using Reddit as a key channel for distribution and community building while ignoring a lawsuit filed in California, according to a report in Ars Technica. The investigation, citing internal documents, indicates that the company has also been acquiring approximately 10 rival services to consolidate its position in the market for creating non-consensual deepfake intimate imagery. This situation highlights the significant jurisdictional challenges authorities face in enforcing laws against illicit AI applications that operate internationally.

Technologists from enterprise companies Atlassian, Intuit, and AWS are actively planning for a future where autonomous AI agents interact directly with software systems via APIs, according to a report in VentureBeat. Based on first-hand experience shared at a 'Women in AI' event, the companies are preparing for a paradigm shift where Application Programming Interfaces must be designed specifically for consumption by AI. This involves creating more standardized and robust APIs to enable AI agents, often powered by Large Language Models, to automate complex business processes. While this promises greater efficiency, it also introduces new IT and cybersecurity challenges that require robust access controls and monitoring systems for agent activity.

Capital One is developing an agentic AI system designed to mirror its own organizational structure to enhance its auto sales processes, according to a report in VentureBeat. The system employs 'agentic AI,' which consists of autonomous agents capable of planning and executing complex, multi-step tasks. Rather than being a single model, the architecture is a collection of specialized AI agents, with each one designed to perform functions analogous to specific departments or roles within the bank. These agents collaborate to manage the entire auto sales workflow, from lead qualification to loan approval, aiming to improve overall efficiency.

A pro-Russia disinformation campaign is leveraging widely available, consumer-grade artificial intelligence tools to significantly increase its output of synthetic content, according to a report in Wired. These easily accessible tools are being used to create a proliferation of fabricated materials, including realistic pictures, videos, QR codes, and fake websites. The use of AI is not creating a new phenomenon but is acting as a powerful amplifier for long-standing disinformation efforts, allowing for the rapid, large-scale generation of diverse and convincing media that is more difficult to detect and counter.

AI-generated videos depicting Black women in a dehumanizing trope as 'Bigfoot' are going viral on platforms like Instagram, according to a report in Wired. Some content creators are reportedly using advanced text-to-video models, such as Google's Veo 3, to produce this content, which is accumulating millions of views. The trend also includes a monetization aspect, with some creators offering to teach others how to make similar videos for a fee.

The social platform X is piloting a program that will use artificial intelligence chatbots to generate drafts for Community Notes, the platform's feature for adding context to posts, according to a report in TechCrunch. This initiative marks an expansion of the system, which was originally launched as 'Birdwatch' by Twitter and has been significantly broadened under Elon Musk's ownership. The current Community Notes system relies on human volunteers to write and rate contextual notes, with a note only becoming visible after achieving consensus from a diverse group of contributors. The new pilot will use AI, likely based on Large Language Models, to create the initial notes, which would then be subject to the existing human review and rating process.

X is introducing a system that allows developers to create AI bots capable of writing Community Notes to provide context on posts, according to a report in The Verge. These AI-generated notes will only be published if they are rated as helpful by human users from diverse perspectives, and they will be clearly marked as originating from an AI. Initially, the AI bots can only write notes on posts where a human has already requested one. The AI writers must earn the ability to contribute and can lose this privilege based on the helpfulness of their notes, mirroring the system for human contributors.

Design platform Figma has filed for an initial public offering (IPO), revealing plans to significantly increase its investment in artificial intelligence, The Verge says in a report. In the filing, CEO Dylan Field stated the company will "double down" on AI, acknowledging it could be a "drag on our efficiency for several years" but is "core to how design workflows will evolve." Figma has already expanded its tools to include AI-powered coding and has started letting AI models access its design servers to make coding more efficient. The move to go public follows a failed acquisition by Adobe for $20 billion, which was terminated in 2023 after facing regulatory pressure.