Google Research

Google has unveiled Project Suncatcher, an initiative to run machine learning infrastructure in space, using solar-powered satellites to house AI compute. The company published details on its research blog, and the announcement is drawing attention among technologists debating the feasibility, cost and environmental logic of moving data centers off Earth.

The European Commission has ordered Google to share anonymized search data, according to a MediaPost report dated July 20, 2026. The move fits the EU's ongoing scrutiny of Google's dominance in search and its efforts to give rivals and researchers access to data. Details of the decision's scope and Google's response are not yet available.

Google researchers say the hacking group ShinyHunters has broadened its attacks targeting Oracle's PeopleSoft software. The group, previously known for large-scale data theft campaigns, is reportedly exploiting vulnerabilities to compromise enterprise systems. The finding raises concerns for organizations running PeopleSoft, with security teams urged to patch systems and review exposure to the campaign.

A Norwegian science news report describes a study finding that children and teenagers aged 9 to 18 increasingly turn to AI assistants instead of Google for information and answers. Researchers say the shift is happening fast, but the effects on how young people learn, judge sources and handle misinformation are not yet known, prompting calls for closer attention from educators and parents.

Google researchers say the hacking group ShinyHunters has broadened its attacks targeting Oracle's PeopleSoft software. The group, known for large-scale data theft and extortion, is reportedly exploiting vulnerabilities to breach organizations using the enterprise platform. The findings raise concerns for companies running PeopleSoft, as security teams are urged to patch systems and monitor for intrusion attempts linked to the campaign.

Google Research has open-sourced RRSI, a framework allowing AI agents to refine their own evaluation harness while guarding against overfitting. Announced via MarkTechPost, the release lets developers inspect and build on the underlying code. The announcement is drawing attention from AI practitioners interested in agent self-improvement methods that remain reliable rather than gaming their own benchmarks.

Google DeepMind has released Gemini 4 Argon, a new AI model offering a context window of up to one million tokens. The upgrade allows the model to process far larger documents, codebases and conversations in a single session than previous versions. Tech observers are weighing the implications for long-form analysis, research and developer workflows, and comparing the context limit with rival AI providers' offerings.

Google DeepMind has introduced SynthID Bio, a set of watermarking methods designed for synthetic biology, extending its SynthID detection technology from AI-generated content into biological sequences. The aim is to embed identifiable markers so synthetic DNA can be traced and verified. The announcement is drawing attention from researchers weighing biosafety, detection reliability and oversight of engineered organisms.

A commentary tagged 'GodsEyeView' argues that layered tools like Google Earth's layers and overlays are more useful than slogans for understanding war, geopolitics and surveillance. The post says these mapping resources have long been valuable to researchers and engineers trying to interpret satellite imagery and geospatial data, and suggests this approach offers deeper insight than headline-level framing of conflicts.

Cybersecurity researchers are flagging a possible phishing campaign hosted through a Google Docs presentation link. The URL, shared in defanged form, points to a public Google Slides page that may be used to lure victims into handing over credentials or personal data. A full technical analysis of the link has been published on a URL scanning service. Users are advised to treat unexpected Google Docs links with caution.

SpaceX launched a Google artificial intelligence satellite along with 129 other payloads, in the second leg of a tripleheader of spaceflight missions. The rideshare flight carried a wide mix of satellites and experiments to orbit, highlighting the growing role of commercial launch providers in deploying large batches of spacecraft, including AI research hardware developed with major tech companies.

Google has published a roundup of the AI news it announced in September 2026, collecting its latest updates on artificial intelligence products and research in one place. The company regularly issues these monthly summaries, and the September edition bundles together its most recent releases for readers tracking its AI roadmap.

Google has introduced SynthID Bio, extending its watermarking technology, originally developed for AI-generated content, into synthetic biology. The tool is designed to embed detectable markers into biological material produced through synthetic processes, helping researchers identify lab-made DNA. The announcement was made via Google's blog, and early reaction is focused on what this means for safety and traceability in biotech.

Cybersecurity researchers are flagging a phishing campaign hosted through a Google Slides sharing link. The URL has been defanged and submitted to a URL analysis service for inspection, and warnings are circulating among infosec accounts urging users not to open such links. The case highlights how attackers continue to abuse trusted Google-hosted pages to deliver scams.

Google has published 'AI in Science: Early Insights', a 42-page report by researchers including Mihai Codreanu, Alex Imas, Juan Mateos-Garcia, Joseph Emmens, Evalyne Muiruri, Arthur Turrell and Julian Jacobs. It looks at how heavily scientists are using AI in their work and how the outputs of that research can in turn be reused to train future AI models, creating a feedback loop between science and machine learning.

Security researchers are flagging a fraudulent Wells Fargo login page hosted on a Google Sites address, designed to steal customers' banking credentials. The link has been defanged to prevent accidental clicks, and a public scan on URLDNA lets others inspect the site's characteristics. The warning has circulated among infosec community members tracking phishing campaigns.

Security researchers report that four separate spyware groups deployed the same exploit kit targeting Google Chrome and Microsoft Windows, all within a single week. The finding suggests the groups are sharing or purchasing identical hacking tools rather than developing their own, raising fresh concerns about the proliferation of surveillance capabilities. Cybersecurity commentators are highlighting the case as evidence of a growing grey market for exploits used against journalists, dissidents and other targets.

Security researchers have flagged two newly published vulnerabilities in Google Android 16 and 16-qpr2. CVE-2026-0016 is a CredentialManager permissions bypass that can expose local information, while CVE-2026-0017 is a BiometricService logic flaw allowing local privilege escalation. No exploitation in the wild has been reported so far. Users and administrators are advised to follow Google's security bulletins for patches.

Security researchers are warning about a phishing site that mimics the Google account sign-in page but is hosted on a raw IP address rather than a Google domain. The fake login page is designed to steal usernames, passwords and two-factor codes from people who click links in scam emails or messages. Users are advised to check that the address is accounts.google.com before entering credentials.

Cybersecurity analysts are warning of a suspected phishing campaign using a malicious Google Docs presentation link. The URL was shared in defanged form on an infosec forum, alongside a scan link for further analysis. Hosted Google Docs presentations are a common phishing vector because they appear legitimate and bypass some email filters. Users are advised to avoid opening unsolicited document links.

Google says artificial intelligence is changing both the speed and the character of vulnerability discovery in cybersecurity, allowing flaws in software to be identified and reported far faster than with traditional methods. According to SecurityWeek, the company argues this acceleration is altering the profile of bugs researchers and defenders encounter, with implications for how organisations patch and prioritise security work.

Cybersecurity researchers are warning about a phishing page hosted on Google Sites, sharing a defanged link and a scan report on urlDNA for analysis. The alerts circulated in infosec channels, with warnings that the site is designed to deceive visitors into handing over credentials or personal data. The use of a legitimate Google domain highlights how attackers exploit trusted hosting services.

Google is testing AI chips in orbit as part of research into data centres based in space. The experiment explores whether satellites equipped with specialised processors could one day handle computing workloads beyond Earth, reducing reliance on land-based data centre infrastructure. Details on the scale, timeline and partners of the project have not been confirmed.

Google DeepMind is extending its SynthID watermarking technology to biology with a tool called SynthID Bio, which embeds hidden markers into AI-designed proteins. The aim is to let researchers identify whether a protein sequence was generated by artificial intelligence, helping distinguish synthetic designs from natural ones and supporting safety and oversight in biotechnology research.

Google has announced a watermarking method for AI-designed proteins, embedding detectable markers into protein structures generated by its models. The goal is to let researchers and regulators identify when a biological tool came from AI design systems. Observers say that if the watermark proves robust, it could become a de facto standard for responsible AI protein design and help distinguish benign research from potentially harmful applications.

Cybersecurity observers are circulating a warning about a fraudulent domain impersonating Google, identified at google29.m4ntapaset.ink. The suspicious link is being shared in defused form so others can avoid clicking it, and a public analysis of the site has been published on the URLdna scanning service. The alert is tagged under cybersecurity, phishing and scam topics as experts urge caution with unfamiliar links.

A critical vulnerability, tracked as CVE-2026-95281, has been disclosed in Google Chrome on Android. The buffer overflow in the ANGLE graphics library, carrying a CVSS score of 9.6, affected Chrome versions before 154.0.8037.57 and could let a remote attacker run arbitrary code outside the sandbox through a crafted HTML page. Security researchers are urging Android users to update Chrome immediately.