Skip to content

Let's Data ScienceLEARN • BUILD • STAY AHEAD

News
Blog
Code Problems
Pricing
Contact

© 2026 Let's Data Science

Advertise|Terms|Privacy||Image Rights

Home
News
Cybersecurity

Topic24 articles · Updated June 4, 2026

AI Cybersecurity News

The intersection of AI and cybersecurity: model vulnerabilities and CVEs, prompt injection and jailbreaks, adversarial attacks on ML systems, and the defensive tooling being built to push back.

executive-order
7.2
Jun 4, 2026
Trump Signs Voluntary AI Model Review Executive Order
President Donald Trump signed an executive order on June 2 titled "Promoting Advanced Artificial Intelligence Innovation and Security," the White House says. The order directs federal agencies to develop benchmarks for assessing AI cybersecurity, create an "AI cybersecurity clearinghouse," and solicit voluntary submission of the most capable or "covered frontier models" to government cybersecurity testing up to 30 days before public release, according to the White House and reporting by Reuters and NPR. The order explicitly avoids imposing a licensing scheme or mandatory preapproval, Reuters and Federal News Network report, but several commentators told Politico the move expands Washington's willingness to scrutinize AI. Supporters and skeptics quoted in Politico and Reuters differ on whether the voluntary framework will become a pathway to stronger rules.
Quick Read+5
crowdstrike
6.8
Jun 4, 2026
CrowdStrike reports rising expenses amid AI investment
CrowdStrike reported a 15% increase in first-quarter operating expenses to $1.07 billion, Reuters reports. The cybersecurity vendor also issued 2027 revenue guidance of $5.91 billion to $5.96 billion, up slightly from prior guidance of $5.87 billion to $5.93 billion, according to Reuters. Total first-quarter revenue was $1.39 billion, versus analysts' average estimate of $1.36 billion, LSEG data cited by Reuters shows. Shares fell about 9% in extended trading, Reuters reports. The company has recently launched Falcon Data Security and the Charlotte AI AgentWorks Ecosystem, and announced a four-for-one stock split, coverage by Reuters and the Economic Times notes.
Quick Read
anthropic
6.7
Jun 3, 2026
Australia Gains Access to Anthropic's Mythos AI
ABC News reports that Australia and several private companies are among 150 organisations across 15 countries granted access to Anthropic's Claude Mythos Preview as part of an expansion of Project Glasswing. ABC reports the model has been withheld from broad public release because of its cybersecurity capabilities; the company previously limited access to UK and US governments and about 50 US companies, ABC reports. The federal government declined to confirm details of its access, while the nation's cyber security agency welcomed the inclusion, ABC reports. Independent MP Kate Chaney is urging the government to accelerate its response to AI risks and opportunities, ABC reports. ABC also cites a new report warning that an AI-driven surge in data centres in Australia threatens the country's energy transition.
Quick Read
anthropic
7.2
Jun 3, 2026
Anthropic Expands Glasswing, Promises Mythos-class Public Release
Per Anthropic's June 2 blog post, Project Glasswing is expanding by about 150 new organizations, bringing the partner list to roughly 200 and covering partners in more than 15 countries. Anthropic reported in the same post that Project Glasswing partners have so far found more than 10,000 high- or critical-severity security flaws. Reuters reported on May 28 that Anthropic intends to make Mythos-class models available "to all customers in the coming weeks" while also launching Claude Opus 4.8. Coverage from TheNextWeb and Reuters notes that Opus 4.8 improves benchmark scores and model "honesty", and that pricing remains unchanged from the prior Opus release. 9to5Mac and Anthropic's announcement say the expanded Glasswing cohort now includes additional major vendors and national partners, with 9to5Mac reporting names such as Samsung and NATO in the wider group.
Quick Read+5
anthropic
6.9
Jun 3, 2026
South Korea Secures Access to Anthropic's Mythos Model
Reuters reports that South Korea's Science Ministry said the Korea Internet & Security Agency (KISA) secured access to Anthropic's cybersecurity AI model Mythos through participation in Project Glasswing. The ministry said the initiative is aimed at using frontier AI models to identify and help fix cybersecurity vulnerabilities. Financial Times reported that Anthropic plans to expand access to about 150 organisations in more than 15 countries, and named Samsung Electronics, SK Hynix, and SK Telecom among participants. Reuters adds that Samsung declined to comment and that SK did not immediately respond. The ministry said South Korea will continue efforts to improve cybersecurity capabilities, including using frontier AI models and strengthening domestic AI-based information security technologies, Reuters reports.
Quick Read
executive-order
7.6
Jun 3, 2026
U.S. asks AI firms to submit models for cybersecurity tests
The White House issued an executive order titled "Promoting Advanced Artificial Intelligence Innovation and Security" that directs federal agencies to seek voluntary agreements with AI developers to submit their most capable models for government cybersecurity testing, according to the order posted on June 2, 2026. The order names the Departments of Treasury, Defense, Commerce and Homeland Security and gives agencies up to 30 days to test models before companies release them externally, per the executive order. Reuters and The New York Times report the timeline was shortened from an earlier 90-day proposal, and the order asks the Treasury secretary to work with banks and critical infrastructure providers on vulnerability scanning, Reuters reports. Industry reaction was mixed; Reuters quotes Google executive Kent Walker calling the order "an important step forward."
Quick Read+7
anthropic
7.3
Jun 3, 2026
Anthropic Expands Mythos Access to 150 Organizations
According to Anthropic's blog post, the company is expanding Project Glasswing access by roughly 150 additional organizations after an initial cohort of about 50 partners tested Claude Mythos. Anthropic reported that early partners detected more than 10,000 high- or critical-severity flaws in their codebases, and Bloomberg reports the total groups with access will be about 200. Per Anthropic and CNBC, the new partners span more than 15 countries and include sectors such as power, water, healthcare, communications, and hardware; each organization must meet security requirements before gaining access. CNBC also reports Anthropic has confidentially submitted a draft S-1 to the SEC and intends to offer Mythos in the European Union.
Quick Read+4
executive-order
7.2
Jun 3, 2026
Trump signs scaled-back AI executive order
President Donald Trump has privately signed a scaled-back executive order creating a voluntary federal review process for powerful AI models, Reuters and Politico report. The order asks developers to provide the government up to 30 days of early access for testing before a public release, down from a previously reported draft that sought 90 days, according to Politico and Reuters. The directive directs agencies to secure voluntary testing agreements and to bolster federal cybersecurity work, Reuters reports, and tasks agencies with creating a benchmarking process and other programs, per SiliconANGLE. The order references covered frontier models and names steps to expand cybersecurity tooling across federal, state, and critical-infrastructure operators, SiliconANGLE and The Guardian report. Editorial analysis: Industry observers should view this as a notable federal step on AI security that remains voluntary and narrower than earlier drafts.
Quick Read+8
executive-order
6.8
Jun 3, 2026
Trump Signs Voluntary AI Model Review Executive Order
President Donald Trump signed an executive order on Tuesday establishing a voluntary review process for advanced AI models, RedState reports. The order creates a category called "covered frontier models" that federal officials will identify through a classified benchmarking process, and it allows developers to voluntarily give the government up to 30-day review before wider release, according to RedState. The order also directs the Cybersecurity and Infrastructure Security Agency (CISA), the Treasury Department, and the NSA to hunt vulnerabilities and push defensive AI tools to agencies, hospitals, banks, and utilities, per RedState. The text explicitly states, "Nothing in this section shall be construed to authorize the creation of a mandatory governmental licensing, preclearance, or permitting requirement," a passage published in the order and cited by RedState. RedState adds that earlier drafts reportedly proposed a 90-day review window.
Quick Read
ai-policy
7.1
Jun 2, 2026
US Executive Order Seeks 30-Day Model Access
President Donald Trump signed an executive order on June 2, 2026, that creates a voluntary framework for federal review of advanced AI systems before public release. Per the White House text, the order asks AI developers to participate in a benchmarking process to determine whether a model is a "covered frontier model" and, on a voluntary basis, to provide government access to such models up to 30 days before broader deployment (White House; CNBC). The order explicitly states it does not create a mandatory licensing, preclearance, or permitting requirement for model releases (White House; CNBC). Reporting from TechCrunch and The Guardian notes the final text is narrower than earlier drafts that sought up to 90 days of pre-release access, and that Trump signed the order privately after delaying a planned public event (TechCrunch; The Guardian).
Quick Read+14
anthropic
6.8
Jun 2, 2026
Canada Gains Access to Anthropic's Mythos Model
Canada's federal government has joined Anthropic's Project Glasswing, giving officials access to the powerful AI model Mythos, AI Minister Evan Solomon told reporters, according to Global News. Solomon said, "I can confirm that the Canadian government is part now of Project Glasswing, which allows companies to have access to Mythos." Anthropic told Global News it launched Project Glasswing to let a small set of trusted organisations and governments use Mythos to scan systems for security vulnerabilities. Anthropic said it expanded the program to 150 additional organisations in more than 15 countries and that partners including Apple, Google, Amazon and Microsoft have found more than 10,000 security flaws they judged highly or critically severe. Anthropic also described Mythos as so "strikingly capable" at hacking and cybersecurity tasks that it withheld the model from general public release.
Quick Read
anthropic
7.8
Jun 2, 2026
Anthropic Expands Project Glasswing to 150 Organizations
According to Anthropic's blog post, the company is expanding Project Glasswing by granting access to its Claude Mythos Preview model to roughly 150 additional organizations across more than 15 countries. Anthropic previously opened the program to about 50 partners, and the expanded cohort includes organizations in power, water, healthcare, communications, and hardware, as well as vendors and nonprofits that maintain widely reused codebases (Anthropic blog). Reporting by TechCrunch and Politico identifies named new participants that include NATO, the EU cybersecurity agency ENISA, and firms such as Okta, Samsung, SK Hynix, and SK Telecom. SiliconANGLE and other outlets cite concerns from security practitioners about controlled rollouts and transparency; SiliconANGLE quotes Justin Beals of Strike Graph warning that opaque access is not a security strategy. Anthropic and other outlets report participating partners have already discovered thousands of vulnerabilities during early testing.
Quick Read+8
palo-alto-networks
7.1
Jun 2, 2026
Palo Alto Networks Raises Guidance After Earnings Beat
CNBC reports that Palo Alto Networks topped Wall Street's third-quarter estimates and issued stronger guidance. According to CNBC, adjusted EPS was $0.85 versus $0.80 expected, and revenue was $3.00 billion versus $2.94 billion expected, a 31% year-over-year increase that included $388 million from the CyberArk and Chronosphere acquisitions. CNBC reports the company posted a net loss of $177 million, or a 22 cent loss per share, versus net income of $262 million a year earlier. CNBC also reports that Palo Alto shares have rallied roughly 60% year-to-date. CEO Nikesh Arora said in a release, "The latest advancements at the AI frontier have increased the level of urgency around cybersecurity, and redefined the shape of the industry for the coming years." Industry context: advances in models like Mythos are cited by reporters as increasing demand for sophisticated cyber defenses.
Quick Read
anthropic
7.8
Jun 2, 2026
Anthropic expands Mythos access to critical infrastructure
Anthropic said in a company blog post that it is expanding Project Glasswing, its controlled program that gives vetted partners access to the Claude Mythos Preview for vulnerability scanning, to roughly 150 new organizations across more than 15 countries. Anthropic wrote that initial Glasswing partners have reported finding more than 10,000 high- or critical-severity flaws, and that the newly added organizations span sectors including power, water, healthcare, communications, and hardware. Reporting by TechCrunch and others, citing the Financial Times, identified some named partners including Okta, Samsung, SK Hynix, SK Telecom, NATO, and the EU cybersecurity agency ENISA. Editorial analysis: the move accelerates real-world testing of high-capability security models while raising oversight and disclosure questions for practitioners.
Quick Read+4
mythos
6.9
May 31, 2026
Bank of England Says UK Banks Lack Mythos Access
Bank of England governor Andrew Bailey told Bloomberg that British banks still have not gained access to Anthropic PBC's Mythos model six weeks after it first drew concern. Bailey said Anthropic was willing to share the model on a trial basis but that access appears to be held up by the US administration, and he called access important given cyber risk, according to Bloomberg and Reuters. Bailey also said banks are using other models to test defenses, and Reuters reported some cybersecurity experts believe fears of unfettered hacking enabled by the model are overstated. Reporting notes that US President Donald Trump postponed signing a broader executive order on AI that had been expected to create a voluntary pre-release engagement framework for developers, per Bloomberg and Economic Times.
Quick Read+1
insurance
6.7
May 29, 2026
Insurers Strengthen Defenses Against AI-Driven Threats
Digital Insurance reports that artificial intelligence is increasing fraud and cyber risk for insurers and prompting a shift in defensive priorities. According to the Information Security Forum, organisations should operate as though a breach is inevitable, define a "minimum viable organization," classify critical data assets, enforce least-privilege access, and patch known vulnerabilities immediately, with compensating controls for unpatchable legacy systems. Digital Insurance cites Ryan Kratz, head of cyber for North America at MSIG, saying AI has made ransomware faster and easier to execute and that concentration among 5-10 major cloud and security vendors creates systemic aggregation risk. Digital Insurance also reports that Capgemini found only 10% of P&C insurers are successfully scaling AI and recommends shifting AI spend toward change management.
Quick Read
sentinelone
6.8
May 29, 2026
SentinelOne Cuts Workforce to Back AI Investments
CNBC reports that SentinelOne shares fell 12% after the company announced plans to cut 8% of its full-time workforce to focus on artificial intelligence and data investments. CNBC reports that CEO Tomer Weingarten told analysts on the earnings call, "This is not a reactive measure, it is a deliberate evolution to reduce complexity, raise the performance bar, and build a leaner, more agile SentinelOne." CNBC reports the company expects a one-time $25 million charge tied to the layoffs and had over 3,000 employees at the end of April. CNBC also reports that SentinelOne issued quarterly revenue guidance of $289 million to $291 million, below the $292 million LSEG consensus, and reiterated full-year revenue guidance of $1.195 billion to $1.205 billion, short of a $1.21 billion forecast.
Quick Read
anthropic
7.1
May 29, 2026
EU intensifies U.S. talks on cyber AI models
The European Commission intends to "intensify" discussions with the U.S. administration on the most advanced AI models, including those with cyber capabilities, a Commission spokesperson told CNBC on May 29, 2026. CNBC reports that Anthropic's Mythos model, announced in April, has prompted concerns about AI-enabled cyberattacks and that Anthropic rolled out the model to a limited preview called Project Glasswing. The Wall Street Journal reported that the White House opposed Anthropic's proposal to expand Mythos access to roughly 70 additional organisations. CNBC also reports that the White House says it is working closely with AI labs to balance innovation and safety, and that OpenAI has offered the EU access to GPT-5.5-Cyber under a limited preview.
Quick Read
india-singapore
6.8
May 29, 2026
India and Singapore Hold 16th Defence Dialogue
India and Singapore held the 16th Defence Policy Dialogue in Singapore, co-chaired by India Defence Secretary Rajesh Kumar Singh and Singapore Permanent Secretary (Defence) Joseph Leong, according to a Ministry of Defence post reported by PTI and The Week. The delegations reviewed bilateral defence cooperation and discussed expanding collaboration in cyber security, artificial intelligence, maritime security, unmanned systems, and advanced defence technologies, per a defence ministry statement reported by Times of India. The Indian delegation visited the Singapore Armed Forces' Digital Operations Technology Centre (DOTC), a visit described by the ministry as providing insights into Singapore's digital and technical capabilities. Both sides reaffirmed support for cooperation through ASEAN-led mechanisms and the ADMM-Plus framework and reiterated commitment to an open, inclusive, rules-based Indo-Pacific, the defence ministry statement said.
Quick Read+6
earnings
7.1
May 28, 2026
Canadian Banks Beat Estimates as IBM, South Korea Invest in AI
StockMarketWatch reports that Royal Bank of Canada posted adjusted EPS of C$3.90 in Q2, above the analyst consensus of C$3.77, and maintained a Basel III CET 1 ratio of 13.5% while setting aside C$912 million for credit loss provisions. StockMarketWatch also reports CIBC beat expectations with adjusted EPS of C$2.54 and said it will sell its 91.61% interest in CIBC Caribbean for $1.6 billion. On AI, StockMarketWatch reports IBM committed $5 billion to a cybersecurity-focused AI initiative, and the South Korean Industry Ministry announced an 800 billion won investment in Furiosa AI. The same report notes mixed corporate results elsewhere, including Hormel Foods missing profit estimates and an EU probe into JD.com's acquisition of Ceconomy.
Quick Read
cybersecurity
6.8
May 28, 2026
Cybersecurity Hiring Surges Amid AI-Driven Tech Layoffs
AI-driven layoffs have continued across the technology sector, with multiple companies including Microsoft, Amazon and Oracle publicly linking cuts to AI adoption, India Today reports. India Today says these reductions have affected tens of thousands of employees and that Meta reportedly cut about 8,000 roles as part of an AI-focused restructuring. Reporting by The New York Times, cited in India Today, finds demand for cybersecurity experts has become so strong that recruiters are struggling to fill open roles. India Today adds that governments, banks and large corporations are increasingly worried advanced AI tools can discover software and infrastructure weaknesses much faster, and that organisations are rushing to bolster security teams. India Today also reports a cybersecurity startup, Calif, used a preview AI tool to test Apple M5 protections in less than a week.
Quick Read
openai
7.0
May 28, 2026
OpenAI Includes South Korea in Cybersecurity Program
Yonhap reports that South Korea's Ministry of Science and ICT held a working-level workshop with OpenAI on May 18, 2026, where Sasha Baker, head of national security policy for ChatGPT, introduced Trusted Access for Cyber (TAC) and related cybersecurity functions of the company's latest models. Aju Press reports that on May 27 Vice Minister Ryu Je-myung met OpenAI Chief Strategy Officer Jason Kwon to formalize Seoul's participation in TAC, and that the Korea Internet & Security Agency (KISA) will serve as the implementing body. Aju Press also reports Korea and Japan are among the first two Asian countries to join the program. Aju Press published a direct quote from Vice Minister Ryu: "Through this cooperation with OpenAI, Korea has laid the groundwork to get ahead of AI-driven cyber threats." Editorial analysis: Industry observers should view this as part of a broader pattern where governments seek vetted, controlled access to frontier models for defensive cyber research and vulnerability assessment.
Quick Read+2
openai-daybreak
7.4
May 27, 2026
OpenAI Launches Daybreak for Cybersecurity Defense
On its Daybreak webpage, OpenAI describes Daybreak as a cybersecurity initiative that combines its GPT-5.5 family with the Codex agentic harness to find, validate, and help remediate software vulnerabilities, accelerate triage, and produce audit-ready evidence, per OpenAI. CyberScoop reports Daybreak is offered in three access tiers, including GPT-5.5-Cyber for controlled red-teaming and a Trusted Access for Cyber tier for verified defensive workflows. Reporting by CyberScoop and AI Business situates Daybreak as a response to Anthropic's Claude Mythos, which Anthropic has kept tightly restricted. Security practitioners and analysts caution these tools could surface far more findings than most organizations can triage; Eric Parizo told TechTarget that vulnerability management programs will face added strain. The Australian Financial Review reports major Australian banks are already using GPT-5.5-Cyber for defensive testing.
Quick Read+7
venture-capital
6.6
May 27, 2026
Seed Investors Shape Early AI And Security Winners
Forbes profiles early-stage investors on the Forbes Midas Seed List, highlighting seed bets that preceded the AI funding surge, according to Forbes. Forbes reports that Gili Raanan and Cyberstarts invested $6.4 million in Wiz's seed round in early 2020, and that Google later acquired Wiz for $32 billion, with Raanan's stake reportedly returning roughly 200 times his original check. Editorial analysis: Early seed capital on frontier AI, cybersecurity and infrastructure disproportionately concentrates upside when category-defining exits emerge, a pattern observers have seen across multiple cycles.
Quick Read

Showing 24 articles

See the full news feed