cybersecurity

A cybersecurity disclosure has placed infrastructure linked to JEE Advanced 2026 under scrutiny after researcher Rylen Anil alleged that a cloud storage configuration exposed thousands of candidate-related records and admit-card PDFs. IIT Roorkee, the organising institute, acknowledged the configuration issue and said corrective action was being taken. The development comes amid recent cybersecurity concerns involving CBSE's On-Screen Marking system and NTA's re-examination portal, highlighting growing attention on data security across India's examination ecosystem.

An IIT Kanpur and IIT Madras expert panel has reportedly found that AI tools, including Claude, were used to identify vulnerabilities in CBSE’s On-Screen Marking (OSM) system, exposing serious security gaps in its digital infrastructure, as per The Economic Times. The panel also flagged shortcomings in vendor Coempt Eduteck’s capabilities, prompting data migration to a government-controlled AWS India setup. MeitY and CERT-In have intensified oversight and audits amid wider concerns over examination cybersecurity and procurement practices.

In a significant policy shift, President Trump has mandated leading AI firms to voluntarily submit their advanced models for government cybersecurity testing. This 30-day pre-release screening aims to identify vulnerabilities and bolster national security, addressing growing concerns over powerful new AI systems. Tech giants like Google and OpenAI are already engaged in discussions with officials.

India's entry into Anthropic's Project Glasswing, gaining access to advanced AI for cybersecurity, promises enhanced threat detection. Experts highlight its potential to revolutionize proactive security measures. This inclusion positions India as a key player in global AI governance and fosters collaborations, though long-term AI competitiveness hinges on infrastructure like energy and silicon.

The National Testing Agency’s (NTA) re-examination portal has come under scrutiny after cybersecurity researchers alleged vulnerabilities that could expose administrative data and provide unauthorized access to portal functions. The claims, shared online, remain independently unverified, and NTA has yet to respond to queries seeking clarification. The allegations emerge amid heightened concerns over the security of examination-related digital platforms following the recent controversy surrounding CBSE’s On-Screen Marking (OSM) system.

Chinese hackers are identified as the primary espionage threat to global tech firms, particularly targeting the booming AI sector for intellectual property and economic secrets. A new report highlights an "AI arms race" with China aiming for global dominance by 2030. Beijing vehemently denies these accusations, calling them political slander while acknowledging ongoing US-China tech discussions.

US government agencies are urgently patching a critical security flaw in Check Point VPN products, identified as CVE-2026-50751. Hackers are actively exploiting this vulnerability to gain unauthorized remote access, with some attacks linked to Qilin ransomware. CISA has mandated fixes by June 11, warning of significant risks to federal networks.

IIT Roorkee has rejected reports of a large-scale JEE (Advanced) 2026 data leak, calling them misleading and factually incorrect. The institute acknowledged a temporary cloud storage misconfiguration on June 2 that exposed limited read-only data but said the issue was quickly fixed. According to IIT Roorkee, no bulk downloads occurred and less than 0.05% of data was accessible. cybersecurity researcher Rylen Anil also clarified that he found a vulnerability but saw no evidence of a mass data leak.

President Trump has signed an executive order for early government access to advanced AI models, aiming to assess cybersecurity risks and protect critical infrastructure. This voluntary collaboration with tech firms allows US agencies to review cutting-edge AI up to 30 days before public release, bolstering national security and combating AI-enabled cyber threats.

India has joined an elite group of nations gaining access to Anthropic's advanced AI cybersecurity model, Mythos, through Project Glasswing. This initiative aims to preempt cyber threats by identifying software vulnerabilities. India's inclusion, alongside key allies, highlights its growing significance in global AI development and its extensive digital infrastructure, offering mutual benefits in the race for cyber defense.

Anthropic has launched Claude Fable 5, a powerful AI model now publicly available. This release follows extensive safety measures to prevent misuse in high-risk fields. Fable 5, designed for complex tasks, is priced competitively and offers enhanced token efficiency, marking a significant step in making advanced AI more accessible.

Google's cybersecurity teams warn of targeted data theft campaigns by the UNC3753 threat cluster against US companies. The group uses vishing and social engineering to gain access, exfiltrate sensitive data, and then extort victims with a three-day deadline to negotiate. They threaten to leak stolen client and proprietary information, causing severe reputational and financial damage.

A former IBM executive has filed a whistleblower lawsuit accusing IBM and AT&T of concealing foreign cyber intrusions from the US government. The lawsuit alleges that despite awareness of significant breaches, including those linked to Chinese hackers, the companies continued to secure federal contracts. The Department of Justice declined to intervene in the case, which remains pending.

AI company Anthropic is expanding its Project Glasswing, granting access to its advanced cybersecurity model, Claude Mythos Preview, to over 15 new countries. This initiative provides selected organizations with tools to identify software vulnerabilities and enhance cyber defenses. Despite the expansion, Anthropic maintains its stance against public release due to concerns about potential misuse and the lack of robust safeguards.

Heads may roll at CBSE following an education ministry crackdown. Irregularities are alleged in a controversial tender for the board's digital evaluation system. Explanations from CBSE have been unsatisfactory. The ministry is investigating the awarding of a contract and security gaps. Experts from IITs are now engaged to strengthen cybersecurity safeguards.

AI leader Anthropic has unveiled Claude Fable 5, a powerful new model accessible to enterprise clients and paid subscribers. This advanced AI excels in software engineering, knowledge work, and research, featuring enhanced safety measures. For specialized tasks, Claude Mythos 5 offers an unrestricted version, though Fable 5 comes with a premium price tag.

CBSE has addressed security vulnerabilities found in its OnMark portal, thanking ethical hackers for their assistance. A cybersecurity team, including experts from government agencies and IITs, has been deployed to strengthen the system. The statement comes after 19-year-old ethical hacker Nisarga Adhikary alleged security flaws in CBSE's digital evaluation ecosystem.

IBM is investing $5 billion in Project Lightwell, a cybersecurity initiative to shield open-source software from advanced AI threats. Driven by concerns over AI models like Anthropic's Mythos, the project mobilizes 20,000 engineers from IBM and Red Hat. Major banks like Goldman Sachs and JPMorgan are already on board to bolster their defenses against emerging vulnerabilities.

CBSE reported that its re-evaluation portal faced multiple cyberattack attempts while processing over 16,000 student submissions for verification and re-evaluation of Class 12 answer sheets. The board said a denial-of-service attack generated nearly 1.5 million hits within two minutes, along with over one lakh unauthorized access attempts. Despite the disruptions, the platform remained operational. The portal, delayed for infrastructure upgrades, will stay open until June 6 with extended session limits and improved stability for applicants.

CrowdStrike CEO George Kurtz warns that unchecked AI spending is creating significant security vulnerabilities, not job losses. He likens the rapid evolution of AI-powered cyber threats to the iPhone's disruptive impact, compressing disruption timelines into months. A new AI model can now automate complex attack chains, making anyone a potential cyberhacker, threatening businesses and national security.

CBSE said it successfully blocked a 3.8 million-packet denial-of-service (DoS) cyberattack targeting its verification and re-evaluation portal while maintaining uninterrupted services for students. Despite the attempted disruption, the portal processed over 56,000 applications for answer-book verification and re-evaluation. The Board said technical teams are enhancing performance and security. Partner banks, including SBI, Bank of Baroda, Indian Bank and Canara Bank, processed thousands of payment transactions as CBSE’s digital systems continue to face public scrutiny.

JEE (Advanced) authorities have clarified that a cloud storage misconfiguration identified by an ethical hacker did not lead to any mass extraction of candidate data. IIT-Roorkee stated the issue, which occurred during technical interventions, was immediately rectified and access was restricted, with no impact on examination records or results.

IIT Roorkee has dismissed allegations of a significant data breach impacting JEE (Advanced) candidates, describing social media narratives as misleading. The institution clarified that a brief misconfiguration in a cloud storage system was swiftly corrected, with access strictly controlled to prevent any large-scale data retrieval.

Google has recently laid off employees within its Cloud division, impacting its Threat Intelligence Group and Mandiant. This move, mirroring broader tech industry trends, sees the company reallocating resources towards artificial intelligence development. Google AI CEO Demis Hassabis, however, believes companies should leverage AI-driven productivity gains to expand, not reduce, their workforce.