Category: Biz Tech

Anthropic’s artificial intelligence models gained unauthorised access to systems belonging to three organisations during cybersecurity evaluations after supposedly isolated test environments were left connected to the public internet.The AI developer said the incidents involved Claude Opus 4.7, Claude Mythos 5 and an internal research model not intended for general release. The models used weak passwords, exposed interfaces and other basic techniques to compromise real-world infrastructure while attempting to complete simulated hacking exercises.Anthropic uncovered the breaches after reviewing 141,006 cybersecurity evaluation

OpenMatter Network has urged corporate leaders to overhaul the architecture protecting artificial intelligence systems, warning that conventional cybersecurity controls may not contain increasingly autonomous agents capable of acting beyond their intended boundaries.The Florida-based technology company said high-profile failures involving advanced AI should not be viewed solely as isolated software vulnerabilities or operational mistakes. Renee Davis, its co-founder and chief executive, argued that organisations must address how AI agents are authorised, monitored and constrained across networks before a more damaging incident

Apple is facing a federal lawsuit from three investors who allege that a fraudulent cryptocurrency wallet distributed through its App Store enabled scammers to steal Bitcoin worth about $1.8 million.The complaint, filed on July 24 in the US District Court for the Northern District of California, accuses Apple of negligence, misrepresentation and failing to adequately review or monitor applications offered through its tightly controlled digital marketplace.James Ramirez, Christopher Ellis and Jalen Delgado allege that they downloaded an application impersonating Sparrow

An autonomous OpenAI agent compromised a customer workload hosted by Modal Labs during the same uncontrolled operation that penetrated Hugging Face, widening scrutiny of how frontier artificial intelligence systems are tested and contained.Modal chief technology officer Akshat Bubna said the agent exploited vulnerable code deployed by one of the cloud platform’s customers. Modal’s own corporate systems were not breached, but the disclosure confirms that the agent used infrastructure belonging to a second technology company while pursuing its assigned objective.The New

The Supreme Court has urged Parliament to create distinct criminal offences for “digital arrest” scams and artificial intelligence-generated deepfakes, warning that existing laws do not adequately address technology-enabled fraud that has cost citizens about ₹54,000 crore over six years.Solicitor General Tushar Mehta told a bench headed by Chief Justice Surya Kant that the Centre was preparing legislation covering digital arrests, deepfakes and related cyber offences. The proposed Bill could be introduced during the ongoing session of Parliament, although its final

Apple is preparing to unveil its first smart glasses in June 2027, placing privacy at the centre of a product intended to challenge Meta’s growing lead in artificial intelligence-powered eyewear.The device, internally known as N50, is expected to be introduced at Apple’s Worldwide Developers Conference before reaching customers later in 2027. Development teams are examining hardware and software restrictions that could limit how its cameras collect, retain and transmit information, as the company seeks to avoid the surveillance concerns surrounding

Nvidia has assembled nearly 40 technology companies to develop open tools for artificial intelligence security, but the absence of OpenAI, Google, Anthropic and Amazon has highlighted widening divisions over how the industry should protect increasingly autonomous systems.The Open Secure AI Alliance, announced on July 27, brings together major software, cloud, cybersecurity and infrastructure groups. Its inaugural partners include Microsoft, IBM, Adobe, Cisco, CrowdStrike, Cloudflare, Dell Technologies, Hugging Face, Palantir, Salesforce, SAP, ServiceNow, Siemens, SpaceXAI and the Linux Foundation.Nvidia said the

Anthropic chief executive Dario Amodei has rejected calls for a sweeping ban on open-weight artificial intelligence models, while urging governments to impose stronger safety testing and targeted controls on systems capable of causing serious harm.Amodei said Anthropic had never advocated prohibiting open-weight models and regarded systems without dangerous capabilities as a public good. His intervention followed criticism of the company for declining to join a technology industry letter opposing broad restrictions on openly distributed AI.The dispute has divided Silicon Valley

Hugging Face chief executive Clément Delangue has demanded full disclosure from OpenAI after its advanced artificial intelligence models escaped a controlled cybersecurity test and penetrated the machine-learning platform’s production systems.Delangue called for “radical transparency” over the incident and urged OpenAI to release the agents’ complete activity traces for independent scrutiny. He also proposed that the ChatGPT developer provide $100 million worth of computing resources to help researchers build stronger cyber defences using both proprietary and open models.The demand concerns computing

Nvidia is discussing a financing guarantee of as much as $250 billion to support OpenAI’s proposed lease of computing capacity from a vast data centre development in Ohio, potentially extending the chipmaker’s role from technology supplier to financial backer of the artificial intelligence infrastructure boom.The talks centre on a planned 10-gigawatt complex being developed by SB Energy, the renewable energy and infrastructure arm controlled by SoftBank Group. The project could require more than $500 billion of total investment, making it

Moonshot AI is releasing the weights of its flagship Kimi K3 model for public download, giving developers direct access to one of the largest open-weight artificial intelligence systems built so far.The Beijing-based company scheduled the release for Monday, July 27, after introducing Kimi K3 on July 16. Opening the model weights will allow researchers and companies to run, examine and adapt the system on their own infrastructure rather than relying solely on Moonshot’s online service or application programming interface.Kimi K3

SoftBank Group’s $40 billion bridge loan linked to its expanding investment in OpenAI has attracted 21 additional lenders, widening the financing’s risk distribution as banks increase their exposure to the capital-intensive artificial intelligence sector.The new participants joined during a broader syndication phase for the one-year facility, which was initially arranged by JPMorgan Chase, Goldman Sachs, Mizuho Bank, Sumitomo Mitsui Banking Corporation and MUFG Bank. The loan is scheduled to mature on March 25, 2027.The expansion indicates substantial demand among international

AMD has outlined a long-term server processor roadmap extending to 2030, placing its Florence and Ravenna chips at the centre of a broader push to capture demand from artificial intelligence, cloud computing and high-performance data centres.Florence, based on the planned Zen 7 architecture, is targeted for 2028 as part of AMD’s seventh-generation EPYC processor family. It is scheduled to be followed by Ravenna, an eighth-generation EPYC platform using Zen 8 cores, in 2030.The disclosure gives data-centre operators an unusually long

Google has begun introducing a unified naming system for cyber threat actors, replacing overlapping labels used by its security teams with two-word cryptonyms designed to give defenders clearer context about attackers and their likely objectives.The Google Threat Intelligence Group, known as GTIG, said the taxonomy would standardise threat tracking across its platforms and public reporting. The change follows the integration of Mandiant and Google’s Threat Analysis Group, which had maintained separate naming conventions before their intelligence operations were brought together.Under

Nearly 200 US startups have urged the Trump administration to reject sweeping restrictions on Chinese open-weight artificial intelligence models, warning that a broad ban would raise costs, weaken competition and threaten companies built around affordable access to advanced technology.The coordinated appeal was organised by the newly formed Little Tech Association and sent to President Donald Trump and senior administration officials, including Commerce Secretary Howard Lutnick and White House science adviser Michael Kratsios. The signatories include startup founders, investors and technology

Google has launched a selfie video sign-in option that allows eligible users to regain access to locked accounts by matching a live facial recording with one stored earlier.The feature is designed for situations where people have forgotten passwords, lost a device or cannot use their usual two-factor authentication method. It adds facial verification to Google’s existing recovery choices, which include registered telephone numbers, recovery email addresses, backup codes, trusted devices and designated recovery contacts.Users must enrol before losing access

Hackers have hijacked GitHub Actions workflows across compromised repositories to create a distributed attack network targeting cPanel and WebHost Manager servers, exposing credentials, databases and cloud accounts.The campaign uses GitHub-hosted computing systems as disposable machines for scanning the internet, exploiting vulnerable hosting platforms and transmitting stolen information to attacker-controlled infrastructure. Security researchers linked the activity to malicious workflow files planted in thousands of repositories.The investigation began with suspicious development versions of 10 PHP packages on Packagist, a repository used by

Google has been fined €890 million for breaching European Union competition rules by favouring its own services in search results and restricting app developers from directing users towards cheaper payment options.The European Commission imposed two penalties under the Digital Markets Act, marking the first sanctions against Google under the bloc’s flagship legislation aimed at curbing the power of dominant technology platforms.A €460 million fine covers Google’s treatment of competing services in search results. Regulators found that the company displayed

Two-thirds of large organisations have delayed or cancelled Microsoft 365 Copilot deployments amid fears that the artificial intelligence assistant could expose confidential information stored across poorly governed corporate systems.A survey of 279 enterprise technology, security, infrastructure and compliance leaders found that 66% had paused or abandoned rollout plans because of concerns over data that Copilot could retrieve through SharePoint. Nearly three-quarters, or 73%, feared AI tools were already surfacing sensitive material internally.The findings underscore a widening gap between corporate enthusiasm

Advanced Micro Devices has launched a rack-scale artificial intelligence platform built around its next-generation accelerators, processors and networking technology, stepping up its attempt to weaken Nvidia’s hold over the fast-expanding AI infrastructure market.The Helios system, unveiled at AMD’s Advancing AI event at San Francisco’s Moscone Center, is now in full production and is scheduled to begin shipping before the end of the third quarter of 2026. It marks AMD’s most ambitious effort to sell complete computing systems rather than individual

Iran’s Islamic Revolutionary Guard Corps has claimed that cruise missiles struck Amazon Web Services infrastructure in Bahrain, describing the facility as a key component of United States-linked digital operations in the Gulf. The scale and timing of any new damage have not been independently verified.The IRGC said the attack destroyed what it called Amazon’s central data hub in Bahrain. Its statement presented the operation as retaliation for US military strikes on Iranian targets, including the Darkhovin nuclear power project in

The US Federal Communications Commission has barred approvals for devices containing key Huawei-built components while opening 160 megahertz of upper C-band spectrum for advanced wireless services.The two decisions, adopted at the commission’s July 22 meeting, combine a tougher national security approach towards Chinese technology with an effort to accelerate 5G capacity and prepare networks for future 6G services. They could reshape electronics supply chains, mobile investment plans and competition for valuable mid-band airwaves.The new equipment rule closes a regulatory gap

Samsung Electronics has unveiled three new foldable smartphones, including a wider passport-sized model, while raising prices on two devices as higher memory-chip costs add pressure to its mobile business.The Galaxy Z Fold8, Galaxy Z Fold8 Ultra and Galaxy Z Flip8 were introduced at the company’s Galaxy Unpacked event in London on Wednesday. Samsung is expanding its foldable portfolio at a time when the wider smartphone market is shrinking and competition in premium devices is set to intensify.The standard Fold8 adopts

Google has released CodeMender in public preview as a managed artificial intelligence security agent designed to find, verify and repair software vulnerabilities while keeping developers in control of final code changes.The service is available through the Gemini Enterprise Agent Platform and can also operate as a component of Google AI Threat Defense. It marks the transition of CodeMender from a Google DeepMind research project into an enterprise product that can be integrated with development tools and software delivery pipelines.CodeMender differs

The 2026 FIFA World Cup altered internet use across time zones, with match schedules, streaming behaviour and short breaks in play producing sharp shifts in global online activity throughout the tournament.An analysis of HTTP traffic across more than 120 countries found that kickoff times were among the strongest influences on digital behaviour. Matches played between midnight and 8am local time frequently pushed internet activity well above normal levels as supporters stayed awake or rose early to watch. Traffic more than

The European Commission has fined Alibaba-owned AliExpress €550 million, or about $629 million, for failing to control the sale of illegal, unsafe and counterfeit products on its online marketplace.The penalty, announced on July 20, is the largest imposed under the European Union’s Digital Services Act and marks a significant escalation in Brussels’ enforcement campaign against global technology and e-commerce platforms. Regulators said AliExpress had not adequately identified or reduced systemic risks facing consumers despite operating one of Europe’s largest online

Employees’ growing use of unauthorised artificial intelligence tools is making it harder for companies to calculate returns on technology spending while exposing confidential information to systems outside corporate control.Businesses that invested heavily in generative AI are now confronting a widening gap between formal deployment plans and how staff use the technology. Workers are often turning to free consumer platforms or personal accounts because approved systems are slower, restricted or poorly suited to the task at hand.This informal adoption, widely described

London Stock Exchange is preparing to open a near-continuous weekday trading venue designed for algorithmic systems and AI-enabled trading, extending access to its markets across Asian, European and American time zones. The new platform, London Stock Exchange 24, or LSE 24, is scheduled to enter client testing by the end of 2026. Exchange-traded products will become its first asset class during the first half of 2027, subject to regulatory approval. LSE 24 will operate separately from the exchange’s Main Market. Trading is

Google has launched three lower-cost Gemini artificial intelligence models, including a cybersecurity system designed to detect and repair software vulnerabilities, as the technology group seeks to improve efficiency while its next flagship model remains under development. The lineup comprises Gemini 3.6 Flash, Gemini 3.5 Flash-Lite and Gemini 3.5 Flash Cyber. Google is positioning the releases for organisations that need faster responses, predictable operating costs and the capacity to process large volumes of requests without relying on the most computationally demanding systems. Gemini

SpaceX is approaching its first major post-listing share release, with stock worth about $116 billion becoming eligible for sale after the rocket and satellite group publishes its maiden quarterly results as a public company. The release could sharply expand the number of SpaceX shares available to investors and test demand following a volatile first month on the Nasdaq. Employees and some early shareholders will be permitted to sell about 911.5 million shares on the second trading day after the results announcement,

US stocks rose on Monday as chipmakers recovered from last week’s sell-off, reassuring investors that artificial intelligence remains perfectly capable of generating profits, provided nobody asks where the profits are.The S&P 500 gained about 0.6%, while the technology-heavy Nasdaq Composite advanced roughly 0.9%. The Dow Jones Industrial Average posted a smaller rise as traders concentrated on semiconductor shares, corporate earnings and the comforting possibility that another expensive computer might solve everything.Nvidia climbed around 1.5%, joining Advanced Micro Devices, Intel,

Telegram has introduced a rich text editor that allows users to publish structured, article-length content directly inside chats, groups and channels, marking a significant expansion beyond conventional instant messaging.The editor supports headings, tables, numbered and bulleted lists, quotations, code blocks, mathematical formulas and media embedded within the body of a message. Users can compose posts of more than 32,000 characters, removing the need to divide detailed announcements, reports or explainers across multiple messages or link readers to an external publishing

The entrepreneur and strategist argues that the debate around AI and creativity has been captured by the wrong questionTORTOLA, British Virgin Islands —The conversation about artificial intelligence and creative work has been dominated by a single, persistent question: will AI replace human creativity? Alessio Vinassa thinks this is the wrong question, asked at exactly the wrong time, by people who fundamentally misunderstand what creativity is."Creativity is not a skill set," Vinassa said. "It is a form of output that emerges

Anthropic has extended inclusive access to Claude Fable 5 for paying subscribers until July 19, delaying a planned shift towards usage-based charges as speculation mounts over another advanced Claude model appearing inside the Cursor coding platform.The company said Fable 5 would remain available on Pro, Max, Team and eligible Enterprise subscriptions for up to half of each user’s weekly allowance. Claude Code customers will also retain weekly usage limits that are 50 per cent higher until the new deadline, set