Newsroom – cybernews.pt

Critical Gravity SMTP Flaw Under Active Attack, Exposing API Keys on 100,000+ WordPress Sites

Newsroom — Sun, 21 Jun 2026 16:59:02 +0000

Threat actors are actively exploiting a serious security vulnerability in the popular Gravity SMTP WordPress plugin, potentially exposing sensitive API keys, OAuth tokens, and email service credentials from more than 100,000 websites worldwide. The vulnerability, tracked as CVE-2026-4020, affects Gravity SMTP versions 2.1.4 and earlier.

According to security researchers, the flaw stems from an improperly secured REST API endpoint that can be accessed without authentication. Attackers can query the endpoint and retrieve a detailed system report containing sensitive information such as API keys, authentication tokens, WordPress configuration data, installed plugins, server details, and database information.

The exposed credentials may include integrations with major email providers such as Amazon Web Services, Google, Mailjet, Resend, and Zoho. If stolen, these credentials could be abused to send spam, launch phishing campaigns, or gain deeper access to affected organizations’ infrastructure.

Security firm Wordfence reported blocking more than 17 million exploitation attempts, including a surge of over 4 million attacks in a single day during early June. Researchers from CrowdSec also observed at least 412 distinct IP addresses attempting to exploit the vulnerability, indicating widespread automated scanning and attack activity.

The vulnerable endpoint, /wp-json/gravitysmtp/v1/tests/mock-data, mistakenly grants access to any visitor because its permission check always returns true. By appending a specific parameter, attackers can force the plugin to generate and expose a large JSON system report containing sensitive configuration data.

The plugin developer, RocketGenius, released Gravity SMTP 2.1.5 to fix the issue. Security experts strongly recommend that website administrators immediately update to the latest version, review server logs for suspicious requests, and rotate any API keys, OAuth tokens, or SMTP credentials that may have been exposed.

Key Facts

Vulnerability: CVE-2026-4020
Affected Plugin: Gravity SMTP for WordPress
Affected Versions: 2.1.4 and earlier
Risk: Exposure of API keys, OAuth tokens, SMTP credentials, and system information
Active Installations: Over 100,000 sites
Patch Available: Version 2.1.5
Observed Exploitation: More than 17 million blocked attack attempts reported by Wordfence

The incident highlights the growing trend of attackers targeting WordPress plugins to harvest credentials and gain access to trusted services, making rapid patching and credential rotation essential for affected website owners.

CISA Issues Urgent Alert as FortiBleed Credential Campaign Compromises 86,644 FortiGate Devices Worldwide

Newsroom — Fri, 19 Jun 2026 16:44:53 +0000

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has issued an urgent warning to organizations using Fortinet FortiGate firewalls after a large-scale cyber campaign known as FortiBleed was found to have compromised 86,644 internet-facing FortiGate devices worldwide. According to recent reports, the operation is linked to Russian-speaking threat actors who have been systematically harvesting and abusing credentials from vulnerable Fortinet systems.

The campaign appears to focus primarily on stolen, cracked, and brute-forced credentials rather than a newly discovered Fortinet software vulnerability. Researchers found large databases containing usernames, email addresses, and passwords associated with tens of thousands of FortiGate firewalls and VPN gateways. Earlier disclosures referenced roughly 74,000 exposed devices, but the number of affected systems has since grown to more than 86,000.

Security researcher Volodymyr “Bob” Diachenko uncovered an exposed archive containing credentials tied to approximately 73,932 Fortinet firewall URLs across 194 countries. Investigations suggest attackers conducted more than a billion credential-guessing attempts against FortiGate VPN systems and harvested authentication data that could later be used to gain access to corporate networks.

According to SOCRadar data cited in reporting, generic administrator accounts and built-in Fortinet system accounts account for a significant share of the compromised credentials. The campaign has affected organizations across government, telecommunications, manufacturing, defense, and private-sector environments worldwide.

What CISA Is Advising

CISA is urging Fortinet customers to immediately:

Terminate all active SSL VPN and administrative sessions.
Reset all administrative and VPN passwords.
Enable phishing-resistant multi-factor authentication (MFA).
Review logs for signs of unauthorized access or lateral movement.
Restrict management interfaces from being accessible over the public internet.
Remove unauthorized accounts and verify administrator privileges.
Use modern password protection methods such as PBKDF2 for credential storage.

Fortinet’s Response

Fortinet stated that the leaked credential data does not appear to result from a new breach of the company itself. The vendor says the dataset is likely a collection of credentials gathered from previous incidents combined with successful brute-force attacks against exposed devices. Fortinet emphasized that organizations following security best practices—such as regularly rotating passwords and enabling MFA—face reduced risk.

Potential Impact

Cybersecurity experts warn that successful compromise of FortiGate appliances can give attackers a foothold at the network edge, allowing them to:

Maintain persistent access.
Monitor network traffic.
Steal additional credentials.
Move laterally into internal systems such as Active Directory environments.
Access sensitive corporate or government data.

The FortiBleed campaign is now considered one of the largest credential-exposure incidents involving Fortinet infrastructure, prompting heightened concern among defenders and government agencies worldwide.

GLM-5.2 Becomes the Highest-Ranked Open-Weights Model on the Artificial Analysis Intelligence Index v4.1

Newsroom — Wed, 17 Jun 2026 15:20:00 +0000

June 2026 — Chinese AI developer Z AI has introduced GLM-5.2, which has emerged as the leading open-weights model on the latest Artificial Analysis Intelligence Index v4.1. The model achieved an Intelligence Index score of 51, placing it ahead of all other publicly available open-weights competitors and reinforcing the growing competitiveness of the open AI ecosystem.

According to benchmark results published by Artificial Analysis, GLM-5.2 currently occupies the top position among open-weights models, surpassing notable rivals including MiniMax‘s MiniMax-M3, DeepSeek‘s DeepSeek V4 Pro, and Moonshot AI‘s Kimi K2.6.

Strong Performance Through Advanced Reasoning

A key factor behind GLM-5.2’s performance is its extensive use of reasoning during inference. Benchmark evaluations indicate that the model generates approximately 43,000 output tokens per task, of which roughly 37,000 tokens are dedicated to reasoning processes. This approach contributes to stronger benchmark results, particularly on complex reasoning and agentic tasks, though it also results in higher computational and token costs compared with some competing models.

Significance for the Open-Weights Ecosystem

The achievement marks another milestone for the open-weights AI community. Over the past year, open models have narrowed the performance gap with proprietary frontier systems at an unprecedented pace. GLM-5.2’s ranking demonstrates that open-weights models are increasingly capable of delivering state-of-the-art performance across reasoning, coding, tool use, and autonomous agent workflows.

The result also highlights the growing influence of Chinese AI laboratories in the global AI landscape. Organizations such as Z AI, DeepSeek, Moonshot AI, and MiniMax have become major contributors to advances in open AI models, accelerating innovation and increasing competition across the sector.

Position Relative to Frontier Proprietary Models

While GLM-5.2 now leads the open-weights category, proprietary frontier systems continue to hold the highest overall positions on intelligence benchmarks. Models such as Claude Opus 4.8 remain ahead on aggregate intelligence measures, indicating that a performance gap still exists between the best open and closed systems.

Nevertheless, the gap has narrowed considerably, and GLM-5.2’s performance underscores the rapid pace at which open-weights models are approaching frontier-level capabilities.

The Great Equalizer: How AI is Reshaping the 2026 World Cup

Newsroom — Thu, 11 Jun 2026 16:31:32 +0000

For decades, the difference between footballing giants and underdog nations came down to more than just talent—it was about resources. The world’s wealthiest federations could afford armies of analysts, proprietary software, and private data labs to pick apart an opponent’s every weakness.

That gap just closed.

At the 2026 FIFA World Cup, FIFA and Lenovo have officially launched Football AI Pro, a generative AI assistant designed to put elite-level tactical intelligence into the hands of every one of the 48 competing nations.

What is Football AI Pro?

Think of it as a world-class assistant coach that never sleeps. Built on FIFA’s specialized “Football Language” model, the tool is trained on millions of historical data points and live match metrics. Coaches from any country—whether it’s a perennial powerhouse like Brazil or a first-time qualifier—can now use the platform to:

Analyze Opponent Patterns: Instantly identify recurring defensive shifts, pressing triggers, or common passing lanes.
Get Tactical Suggestions: Receive data-backed insights on substitution windows and tactical adjustments based on real-time fatigue markers.
Visualize the Game: Generate 3D tactical models that make complex formations easy to understand, even for staff with limited resources.

Why This Matters

“We will democratize access to data,” says FIFA President Gianni Infantino. By providing a common, high-level analytical engine to all teams, the tournament is shifting the focus back to the human element of the game: tactical ingenuity.

It’s no longer about who has the biggest budget for private software; it’s about which coaching staff uses these insights to make the smartest decisions under pressure.

The Human Side of Tech

This isn’t just about cold numbers; it’s about the narrative of the game. For the fans, this technology extends to the broadcast. You’ll see AI-smoothed “Referee View” footage and highly accurate 3D player avatars during VAR checks. These innovations are designed to make the “invisible” work of officiating transparent, turning controversial decisions into clear, understandable moments for all 6 billion viewers.

NVIDIA Unveils RTX Spark: 128GB Unified Memory, 6,144 CUDA Cores

Newsroom — Sat, 06 Jun 2026 16:48:42 +0000

NVIDIA has officially entered the Windows PC processor market with the launch of RTX Spark, a powerful new Arm-based system-on-chip (SoC) designed for next-generation AI PCs, laptops, and compact desktops. The platform combines a high-performance CPU, advanced Blackwell GPU architecture, and massive unified memory into a single chip package.

NVIDIA claims the chip can deliver up to 1 petaflop of AI computing performance, enabling users to run large AI models locally, render complex 3D scenes, edit 12K video, and perform advanced creative tasks without relying heavily on cloud infrastructure.

The company is positioning RTX Spark as a major step toward “AI-native” Windows PCs. According to NVIDIA, the platform is designed to support AI agents that can automate workflows, assist with software development, accelerate creative projects, and enhance productivity directly on the device. Microsoft is working closely with NVIDIA to optimize Windows for the new architecture and improve support for AI-powered applications.

For gamers, NVIDIA says RTX Spark-powered systems will offer graphics performance comparable to high-end mobile GPUs, with support for RTX technologies such as ray tracing, DLSS, Reflex, and CUDA acceleration. The company has also highlighted the chip’s ability to deliver AAA gaming experiences while maintaining strong battery efficiency in thin-and-light laptops.

Major PC manufacturers including Microsoft, Dell, HP, Lenovo, ASUS, MSI, Acer, and Gigabyte are expected to launch RTX Spark-based devices later this year. Initial products will target premium laptops and compact desktop PCs, with broader adoption likely to follow as the Windows-on-Arm ecosystem continues to mature.

Key Specifications

CPU: 20-core NVIDIA Grace (Arm-based)
GPU: Up to 6,144 Blackwell CUDA cores
Memory: Up to 128GB unified LPDDR5X
AI Performance: Up to 1 petaflop
Target Devices: Windows laptops and mini PCs
Launch Window: Fall 2026

Industry analysts view RTX Spark as NVIDIA’s most ambitious move yet into the PC processor market, placing the company in direct competition with Intel, AMD, Apple Silicon, and Qualcomm in the race to power the next generation of AI-focused personal computers.

NASA Lifts Evacuation Alert, Astronauts Return to ISS After Air-Leak Repair Effort Paused

Newsroom — Fri, 05 Jun 2026 15:58:30 +0000

Five astronauts aboard the International Space Station were told by NASA to return to normal duties after spending about two hours sheltered inside a docked spacecraft during concerns over a worsening air leak in the station’s Russian segment.

Earlier on Friday, NASA instructed the four-member Crew-12 team and another U.S. astronaut to enter their docked SpaceX Crew Dragon capsule and put on spacesuits as a precaution. The move was part of NASA’s “safe-haven” protocol, used when there is a potential threat to crew safety.

The alert was triggered while Russian cosmonauts were attempting repairs in the aging Zvezda service module, where a long-running air-leak problem has persisted for years. NASA said the leak rate unexpectedly increased, prompting concern that the situation could worsen during repair work.

After further assessment by NASA and the Russian space agency Roscosmos, the shelter order was lifted and the astronauts were allowed to return to the station.

Roscosmos later reported that two leaks had been detected in the transition chamber of the Zvezda module. One leak was successfully sealed with a special compound, while preparations continue to repair the second. The agency said the station’s internal pressure remains stable and there is no immediate danger to the crew.

The incident highlights ongoing concerns about the aging ISS infrastructure. Small leaks in the Russian segment have been monitored since 2019, leading to repeated inspections and repairs. Although Friday’s event prompted a rare evacuation-ready posture, the station has never required a full emergency evacuation in its 27-year history.

5 AI Models Dominating the Industry Right Now

Newsroom — Thu, 28 May 2026 13:49:44 +0000

Artificial intelligence is no longer a future trend — it has become the operating system of modern work. From coding assistants and research agents to multimodal content generation, today’s leading AI models are transforming how businesses, developers, and creators operate.

Here are the five AI models setting the standard in 2026.

1. OpenAI GPT-5.5

The Most Complete AI System Available Today

OpenAI’s GPT-5.5 continues to lead the market as the most balanced and capable frontier model currently available. It combines advanced reasoning, strong coding performance, multimodal understanding, and agent-style task execution into a single platform.

Why it stands out

Exceptional reasoning and analytical depth
Industry-leading coding assistance
Powerful automation and AI agent workflows
Reliable performance across business and creative tasks

Best suited for

Software development
Research and analysis
Enterprise productivity
AI-powered automation

2. Anthropic Claude Opus 4.7

The Leader in Long-Context Intelligence

Anthropic’s Claude Opus 4.7 has become the preferred choice for users who need deep reasoning across extremely large documents and complex workflows.

Core strengths

Massive context-window capabilities
Strong technical writing and summarization
Reduced hallucination rates
Excellent structured reasoning

Best suited for

Legal and research workflows
Document-heavy analysis
Technical writing
Enterprise knowledge management

3. Google Gemini 3.5 Pro

The Most Advanced Multimodal Ecosystem

Google has aggressively expanded Gemini into one of the most versatile AI ecosystems in the market, integrating advanced multimodal capabilities directly into its products and services.

What makes Gemini important

Native image, video, and voice understanding
Deep integration with Google Workspace
Fast “Flash” variants for real-time interaction
Strong search and productivity capabilities

Best suited for

Productivity workflows
Education and research
Content creation
Multimedia analysis

4. xAI Grok 4

The Fastest-Rising Challenger in AI

Originally viewed as an experimental competitor, Grok 4 has evolved into one of the most talked-about AI systems of the year.

Why Grok is gaining momentum

Rapid response performance
Competitive reasoning improvements
Strong coding capabilities
Tight real-time information integration

Best suited for

Real-time interaction
Technical workflows
Fast-paced productivity
Social and information-driven environments

5. DeepSeek V4 Pro

The Open-Model Disruptor

DeepSeek has emerged as one of the most influential open-model players in the AI industry, offering high-level performance at significantly lower operating costs.

Why developers are paying attention

Strong coding and reasoning performance
Cost-efficient API access
Open-weight flexibility
Growing developer ecosystem

Best suited for

AI startups
Self-hosted AI solutions
Independent developers
Cost-conscious businesses

The broader trend

The most effective AI users today are no longer relying on a single model. Instead, they are combining multiple systems depending on the task — using different AI platforms for reasoning, coding, content generation, automation, and research.

The AI era is no longer about experimenting with tools.
It is about building workflows around them.

The AI Revolution: Why the Internet Will Never Be the Same

Newsroom — Tue, 26 May 2026 17:09:41 +0000

Artificial intelligence (AI) is rapidly transforming the internet, changing how people search for information, create content, shop online, and interact with digital services. What was once considered a futuristic technology has become a central part of everyday online experiences, affecting billions of users worldwide.

One of the biggest changes is occurring in online search. Traditional search engines have long relied on displaying lists of links for users to explore. Today, AI-powered systems can generate direct answers, summarize complex topics, and provide personalized recommendations in seconds. This shift is changing how people access information and how websites attract visitors.

Content creation has also been revolutionized. AI tools can now generate articles, images, videos, music, and computer code with remarkable speed. Businesses are using these technologies to improve productivity, while creators are finding new ways to produce and distribute content. However, concerns about misinformation, copyright issues, and the authenticity of AI-generated material continue to grow.

The e-commerce industry is experiencing significant changes as well. AI helps online retailers understand customer preferences, recommend products, optimize pricing, and improve customer support through advanced chatbots. These capabilities are creating more personalized shopping experiences and helping companies increase efficiency.

Social media platforms are increasingly relying on AI to recommend content, moderate harmful material, and enhance user engagement. While these systems can improve user experiences, experts warn that they may also contribute to the spread of misinformation, echo chambers, and privacy concerns if not carefully managed.

Businesses across industries are investing heavily in AI technologies. Major technology companies are competing to develop more powerful AI models and integrate them into their products and services. This competition is driving innovation but also raising questions about regulation, transparency, and the future of work.

Education, healthcare, finance, and customer service are among the sectors benefiting from AI-driven tools. From personalized learning platforms to advanced medical diagnostics, AI is helping organizations solve problems faster and more effectively than ever before.

Despite its benefits, the rapid adoption of AI presents challenges. Governments and regulators around the world are working to establish rules that address issues such as data privacy, algorithmic bias, security risks, and the impact of automation on employment.

Experts believe the internet is entering a new era in which AI will become deeply embedded in nearly every online activity. While the long-term effects remain uncertain, one thing is clear: artificial intelligence is not just changing the internet—it is fundamentally reshaping the way people interact with the digital world.

AI is transforming search, content creation, e-commerce, social media, and business operations, making it one of the most significant technological shifts since the birth of the modern internet.

Viral Gen Z-Led Cockroach Janta Party Founder Alleges Website Block, Online Crackdown and Threats

Newsroom — Mon, 25 May 2026 15:59:34 +0000

The founder of the viral online movement Abhijeet Dipke has alleged that his group is facing a coordinated crackdown after gaining massive popularity among India’s youth. According to Dipke, the movement’s website was blocked, its social media accounts faced restrictions, and members of his family received threats. The allegations have not been independently verified, and the Indian government has not publicly confirmed involvement.

The movement, known as the Cockroach Janta Party (CJP), emerged as a satirical online campaign highlighting concerns such as unemployment, inflation, exam paper leaks, media freedom, and governance issues affecting young Indians. Within days, the group’s Instagram following surged past 20 million, making it one of the fastest-growing political-satire movements on Indian social media.

Dipke claims the group’s official website was taken down and that attempts were made to compromise its social media presence. He also said the movement’s X account was withheld in India and that affiliated accounts faced restrictions.

The controversy has now moved into the legal arena. Dipke has approached the Delhi High Court, challenging the blocking of social media accounts associated with the movement and seeking judicial intervention. The case has sparked debate about freedom of expression, online satire, and digital censorship in India.

The Cockroach Janta Party was reportedly inspired by public outrage after comments comparing some unemployed youth to “cockroaches.” The movement adopted the term as a symbol of youth frustration and resilience, attracting hundreds of thousands of supporters and sign-ups within days.

Google Set to Showcase Major AI Push for Developers and Consumers at I/O 2026

Newsroom — Tue, 19 May 2026 17:03:22 +0000

Google is expected to place artificial intelligence at the center of its annual Google I/O conference, with announcements aimed at both software developers and everyday consumers.

The event, taking place in Mountain View, California, comes as Google intensifies competition with rivals such as OpenAI and Anthropic in the rapidly evolving AI market. According to reports, CEO Sundar Pichai is expected to unveil upgraded Gemini AI models, smarter AI agents, and deeper AI integration across products including Gmail, Chrome, YouTube, Android, and Search.

What Google Is Expected to Announce

Analysts and media reports suggest several major themes for the conference:

New Gemini AI upgrades with stronger reasoning and automation abilities.
Expanded “agentic AI” tools that can complete tasks on behalf of users.
AI-powered coding tools for developers, following Google’s aggressive push into AI-assisted software engineering.
More AI features integrated directly into Google Search and Android.
Possible demonstrations of Android XR smart glasses and AI-enhanced hardware.

Focus on Developers

A major part of the conference is expected to target programmers and enterprise users. Google reportedly wants to strengthen its position in AI coding tools after investing heavily in AI infrastructure and developer products. Reuters noted that Google recently acquired talent from coding startup Windsurf and may reveal new coding assistant technology under projects such as “Antigravity.”

Consumer AI Strategy

Google is also trying to make AI more mainstream for ordinary users. The company has already expanded Gemini into Search and Workspace products and is expected to continue pushing AI experiences that can answer questions, automate tasks, and personalize services. Industry analysts say Google’s huge ecosystem — including Android, Chrome, Gmail, and YouTube — gives it a major advantage in distributing AI tools globally.

Why This Conference Matters

The 2026 I/O conference is seen as especially important because it reflects Google’s effort to prove it can lead the AI race after initially appearing behind ChatGPT’s rapid rise. The company is expected to highlight not only powerful AI models but also practical ways AI can be embedded into daily digital life.