• Search for:
    13Aug, 2024
    Criminal IP and Maltego Collaborate to Broaden Threat Intelligence Data Search

    Criminal IP, an expanding Cyber Threat Intelligence (CTI) search engine from AI SPERA, has recently completed its technology integration with Maltego, a global all-in-one investigation platform that specializes in visualized analysis of combined cyber data.

    This collaboration integrates Criminal IP’s comprehensive database of malicious IPs, domains, and CVEs directly into Maltego’s unified user interface and adds Criminal IP to Maltego’s marketplace, Transform Hub.

    Maltego translates Criminal IP data into a visual data graph, allowing users to easily recognize relationships between each entity and associated risks by adjusting the layouts and assigning weights to them.

     Criminal IP’s C2 tag and vulnerability data visualized through Maltego

    Criminal IP-Maltego Introduces New Features: Visualizing, Tracking Cyber Threat Information

    Now through its partnership with Criminal IP, its trusted data source, and an OSINT CTI tool, Maltego users can also harness Criminal IP’s comprehensive threat intelligence search functionalities to instantly visualize data.

    New key features in Maltego allow users to visualize vulnerabilities by importing Criminal IP’s comprehensive data, including CVEs, assets’ reputation, botnets, Command & Control servers (C2), domain phishing information, and more.

    They can also track exposed personal information in banner data, such as API keys, token values, bank account numbers, and Bitcoin wallet addresses, ensuring prompt identification.

    The tool visually verifies relationships between IP addresses and domains, facilitating rapid response and effective threat tracking.

    Criminal IP’s Domain and IP data visualized through Maltego 

    Maltego: Speeding up Cyber Investigations in a Single Interface

    Maltego is an integration platform with a high impact on the field of threat intelligence and has integrations with several well-known products, including Microsoft Sentinel, IBM QRadar, and Google Maps Geocoding.

    Its existing features drastically accelerate complex cyber investigation by enabling quick preliminary OSINT investigations for digital profiling with Maltego Search as well as complex link analysis for large datasets with Maltego Graph.

    Through Maltego Evidence and Maltego Monitor, the platform enables investigators to collect, monitor, and preserve social media intelligence in real time for prosecution and public safety.

    About AI SPERA

    AI SPERA, renowned for its advanced solutions, has expanded internationally with ‘Criminal IP’ as its flagship offering. Operating in 150+ countries, ‘Criminal IP’ is backed by enterprise-grade security solutions like ‘Criminal IP ASM’ and ‘Criminal IP FDS’. Strategic partnerships with global leaders such as Cisco, VirusTotal, and Quad9 have significantly enhanced ‘Criminal IP’s capabilities. Recently, AI SPERA’s ‘Criminal IP’ has entered the marketplace of major US data warehousing platforms including Amazon Web Services (AWS), Microsoft Azure, and Snowflake, expanding its global reach for threat data.

    Ref: Link

    12Aug, 2024
    LMI buys 3D robot guidance firm, Liberty Robotics
    Liberty Robotics’ VFix technology guides automotive robots using 3D vision to measure six degrees of freedom for large parts or vehicle bodies.

    The Canadian 3D scanning and inspection developer LMI Technologies is buying the US robotic guidance specialist Liberty Robotics for an undisclosed sum. Michigan-based Liberty (formerly known as Liberty Reach), develops 3D machine vision systems to guide robots in automotive, material-handling/packaging, and logistics applications.

    The acquisition will allow LMI to use Liberty Robotics’ strong foothold in North America to expand its global footprint and accelerate its growth in Asia-Pacific and Europe.

    “This acquisition marks a significant milestone in our strategic growth, expanding our solution-based capabilities in the automotive and packaging and logistics industries,” says LMI Technologies CEO, Mark Radford. “Liberty Robotics brings deep expertise in robot integration with large field-of-view snapshot scanning as well as powerful application-level software, including its industry-leading solutions for vision-based robotic guidance and mixed-case palletization that perfectly complement LMI’s advanced smart 3D sensor technology infrastructure and global market presence.”

    Liberty Robotics’ VFix and VGuide systems allow precise end-of-arm robotic guidance for automotive applications such as parts handling and applying coatings and sealers. Its customers include BMW, Tesla, VW, Honda, Toyota, Mercedes-Benz, Nissan, and Stellantis.

    Liberty has used its experience in the automotive industry to develop vision guidance systems for robotic applications in the materials-handling/packaging and logistics sectors. Its VPack and VPick systems support automated mono and mixed-case palletizing, depalletizing, decanting, and delayering for autonomous material handling in logistics centers.

    They can handle mixed pallets of boxes that vary in size, color, and pattern. This will help to drive LMI’s growth in materials-handling applications.

    Unlike contact-based measurements or 2D vision systems, Liberty’s non-contact technologies add 3D shape information for 100% quality control. It also offers AI-based systems that provide deep-learning-driven visual inspection for production processes.

    “We see tremendous synergies between LMI and Liberty Robotics,” says Radford. “Liberty Robotics’ proven technologies and talented team will greatly enhance our ability to deliver integrated, high-value solutions to our customers worldwide. We aim to set new benchmarks in factory automation efficiency and innovation.”

    Ref: Link

    3Aug, 2024
    Google Adds Lens, and Other AI Tools to Chrome on Desktop

    Google Chrome users on Mac, PC, or Chromebook will soon see some new features designed to make searching the web more flexible. The new features are enabled by Google Gemini and other Google AI models.

    Google Chrome’s last major AI update was in May , when it announced Gemini functionality across Workspace and added AI-generated answers to some Search queries.

    What are Chrome’s new features?

    The three new features are:

    • Google Lens in Chrome on the desktop browser.
    • Tab compare, an automatically generated comparison tool for online shopping.
    • Natural language search in browser history.

    Google Lens allows users to select specific elements from a picture or web page. It’s best suited for shopping since it can tell you where to purchase an object in a photo. Google Lens answers general queries, too, such as searching the web for a certain type of plant or more information about an equation. Google Lens in Chrome will be accessible to click-and-drag from the address bar or in the right-click or meatball menu.

    Tab compare generates a chart listing the attributes of two or more products. Users can create a comparison table for products from different marketplaces and vendors, as long as they line those products up in tabs with each other.

    “Online comparison is something we hear users want help with,” wrote Parisa Tabriz, vice president of Chrome at Google, in a press release.

    Perhaps the most remarkable Chrome update from a security perspective is the ability to search your browsing history using natural language. Google suggests a consumer use for this too — the question “What was that ice cream shop I looked at last week?” — but it could have business applications too, such as searching across multiple meetings and communication platforms.

    Cautionary features in place around search history include:

    • Using the AI search feature is optional.
    • The AI search feature can be turned off in Chrome settings.
    • The AI search feature never includes browsing data from incognito mode.

    Tabriz said Google is leveraging AI to make it easier for users to “quickly get things done and find the information you want. Our goal is to keep making your browsing experience even more helpful, through the power of AI,” she explained.

    Ref: Link

    Akchara

    31Jul, 2024
    Surging data breach disruption drives costs to record highs

    Security teams are getting better at detecting and responding to breach incursions, but attackers are inflicting greater pain on organizations’ bottom lines. IBM’s recent Cost of a Data Breach Report 2024 found the global average breach hit a record $4.88 million. That’s a 10% increase from 2023 and the largest spike since the pandemic.

    While the study notes that organizations, on average, improved their time to identify and contain breaches, rising business costs drove the global average breach cost higher. Among the largest contributors were lost business costs, expenses from post-breach customer support (such as setting up help desks and credit monitoring services), and paying regulatory fines. Some 70% of the 604 organizations studied reported that their operations were either significantly or moderately disrupted.

    The new research, conducted independently by Ponemon Institute and analyzed by IBM, studied breached organizations from 16 countries and regions and across 17 industries. It also included interviews with 3,556 security and business professionals from the breached organizations. In its 19th year, the Cost of a Data Breach Report provides actionable insights and up-to-date research, making it a critical benchmark for the industry.

    While the report’s findings suggest some damages from a breach are unavoidable, they also highlight several risk areas that security teams can and should address. For instance, the findings underscore the growing importance of security AI and automation technologies for mitigating breach impacts and lowering costs associated with those breaches.

    Below are those takeaways and several others from the Cost of a Data Breach Report 2024.

    AI and automation in security most effective at reducing average costs

    More organizations are adopting AI and automation in their security operations, up 10% from the 2023 report. And most promising, the use of AI in prevention workflows had the highest impact in the study, reducing the average cost of a breach by $2.2 million, compared to organizations that didn’t deploy AI in prevention.

    Two out of three organizations in the study deployed AI and automation technologies across their security operations center. This factor may also have contributed to the overall decrease in average response times – those using AI and automation saw their time to identify and contain a breach lowered by nearly 100 days on average.

    Only 20% of organizations said they are using gen AI security tools, yet those that did saw a positive impact, with gen AI security tools shown to mitigate the average cost of a breach by more than $167,000. Read the report

    Security staffing shortages led to higher breach costs and more security investment

    Staffing shortages in security departments continued to grow, with 53% of organizations facing a high-level skills shortage, up 26% from 2023. The industry-wide skills shortage could be expensive for organizations. Those with severe staffing shortages experienced breach costs that were $1.76 million higher on average than those with low-level or no security staffing issues.

    These staffing shortages may be contributing to the increasing use of security AI and automation, which has been shown to reduce data breach costs. At the same time, staffing shortages may see some ease, as businesses reported they intend to increase security investments as a result of the breach. Organizations planned investments including threat detection and response tools like SIEM, SOAR, and EDR, according to the report. Organizations also plan to increase investments in identity access management, and data protection tools.

    These additional investments could pay off in mitigating future breach costs. More organizations in 2024 identified the breach with their own security teams and tools (42%) compared to last year (33%), and those organizations had lower than average breach costs, including nearly $1 million lower on average than breaches that were identified by the attacker, such as in an extortion attack.

    Cloud and data security issues remained prominent

    Forty percent of breaches involved data stored across multiple environments including public cloud, private cloud, and on-premise. These multi-environment breaches cost more than $5 million on average and took the longest to identify and contain (283 days), highlighting the challenge of tracking and safeguarding data, including shadow data, and data in AI workloads, which can be unencrypted.

    The types of data records stolen in these breaches underscored the growing importance of protecting an organization’s most sensitive data, including customer personal identifying information (PII) data, employee PII, and intellectual property (IP). Costs associated with customer PII and employee PII records were the highest on average.

    Customer PII was involved in more breaches than any other type of record (46% of breaches). However, IP may grow even more accessible as gen AI initiatives bring this data out in the open. With critical data becoming more dynamic and available across environments, businesses will need to assess the specific risks of each data type and their applicable security and access controls.

    What else is new in the 2024 Cost of a Data Breach Report

    Each year poses new data security challenges as threats and technologies emerge, and this report evolved to reflect these changes. New research conducted for the first time this year in the 2024 Cost of a Data Breach Report included:

    • Organizations experiencing long-term operational disruption, and the time it takes to restore data, systems, or services to their pre-breach state
    • To what extent organizations are using AI and automation in each of four areas of security operations: prevention, detection, investigation, and response
    • How long it took organizations to report the breach if they were mandated to do so
    • Whether organizations that involved law enforcement following a ransomware attack paid the ransom

    Of course, the report continues to showcase the top costliest geographies and industries, the initial causes of data breaches and their costs, and much more. Importantly, the report continues to provide recommendations from IBM experts, addressing the report findings, to help organizations understand the risks and how to mitigate the impacts and potential costs of a data breach.

    Ref: Link

    26Jul, 2024
    Researchers Claim Anyone Can Access Deleted, Private GitHub Repository Data

    Cybersecurity researchers have revealed a GitHub design flaw that allows access to deleted and private repository data. Learn how the issue, dubbed Cross Fork Object Reference (CFOR), puts sensitive information at risk and what it means for open-source security.

    Cybersecurity researchers at Truffle Security, an open-source security software company, have found that anyone can access deleted and private repository data on GitHub, and it’s available forever.

    This issue, known as Cross Fork Object Reference (CFOR), allows users to directly access commit data from another fork, including data from private and deleted forks.

    According to Truffle Security, this is not a bug, but rather an intentional design feature of GitHub’s repository architecture. This means that any code committed to a public repository may be accessible forever, even if the original repository is deleted, as long as there is at least one fork of that repository.

    The researchers demonstrated three scenarios where this issue can be exploited:

    Accessing Deleted Fork Data: When a user forks a public repository, commits code to it, and then deletes the fork, the committed code is still accessible forever. This is because GitHub stores repositories and forks in a repository network, and deleting a fork does not delete the committed data.

    Accessing Deleted Repo Data: When a public repository is deleted, GitHub reassigns the root node role to one of the downstream forks. This means that all commits from the original repository still exist and are accessible via any fork, even if the fork never synced with the original repository.

    Accessing Private Repo Data: When a private repository is open-sourced, any code committed between the time the internal fork was created and when the repository was made public is accessible on the public repository. This is because the private fork and public repository are part of the same repository network.

    To access the data, an attacker only needs to know the commit hash https://github.com/<user/org>/<repo>/commit/<commit_hash>, which can be brute-forced through GitHub’s UI or obtained through the public events API endpoint. This means that confidential data and secrets may be inadvertently exposed on an organization’s public GitHub repositories.

    Truffle Security has submitted their report titled “Anyone can Access Deleted and Private Repository Data on GitHub” to GitHub via their Vulnerability Disclosure Program, and GitHub has responded that their architecture is designed to work this way. While GitHub is transparent about their architecture, the average user may not understand the implications of this design, and the act of deletion does not necessarily mean that the data is deleted.

    Truffle Security recommends that users take steps to securely remediate leaked keys on public GitHub repositories through key rotation. They also note that secret scanning tools will need to be updated to handle these scenarios and that users should be aware that deleted data may still be accessible.

    This issue is not unique to GitHub, and other version control system products may also be affected. As the use of open-source software continues to grow, users must understand the security implications of how these systems work.

    25Jul, 2024
    Chinese SMS Phishing Group Hits iPhone Users in India Post Scam

    The notorious Chinese Smishing Triad gang, known for its SMS phishing attacks against Pakistan, the US, and European nations, has now set its sights on iPhone users in India. The group is exploiting iMessage and the government-owned India Post in a sophisticated phishing scam.

    FortiGuard Labs has revealed a sophisticated Smishing (SMS Phishing) campaign targeting users of India Post, the country’s government-operated postal system. The scam, attributed to a China-based threat actor known as the Smishing Triad, involves sending deceptive iMessages to iPhone users, claiming that a package is waiting for them at an India Post warehouse.

    The fraudulent messages often contain a short URL leading to a fake website designed to mimic the official India Post site. Victims are then prompted to provide sensitive personal information, such as their name, residential address, email ID, and phone number. In some cases, the scammers even request credit card details under the guise of a small redelivery fee.

    FortiGuard Labs’ investigation revealed that between January and July 2024, over 470 domain names were registered to impersonate India Post’s official domain. Notably, 296 of these domains were registered through a Chinese registrar, Beijing Lanhai Jiye Technology Co., Ltd., raising concerns about the intentions behind the campaign.

    The Smishing Triad has previously targeted other regions, including the US, UK, EU, UAE, KSA, and Pakistan. Their modus operandi involves using third-party email addresses, such as Hotmail, Gmail, or Yahoo, to create Apple IDs and send phishing messages via iMessage. This tactic allows the scammers to bypass traditional email security measures and reach users directly on their iPhones.

    According to Fortinet Labs’ report shared with Hackread.com ahead of its publication on Thursday, the phishing campaign is quite sophisticated and well planned. The investment in registering the malicious domain names alone exceeds USD 1500.

    Jason Soroko, Senior Vice President of Product at Sectigo, commented on the issue, stating, “The use of third-party email addresses on iMessage facilitates these attacks, highlighting a need for increased awareness and robust security measures among users to mitigate potential financial losses and data breaches.”

    Stephen Kowski, Field CTO at SlashNext Email Security+, emphasized the need for comprehensive mobile web threat protection, stating, “As smishing attacks become increasingly sophisticated, organizations must prioritize educating their users on how to identify and report suspicious messages, while also implementing robust security measures that can inspect and mitigate threats in real-time, regardless of the communication channel used.”

    To protect themselves from such scams, users are advised to be cautious of unexpected emails or messages, verify URLs before clicking on them, and avoid sharing personal information via email or messaging apps. Enabling multi-factor authentication and keeping software up to date can also help strengthen account security.

    Ref:https://hackread.com/chinese-sms-phishing-group-iphone-users-india-post-scam/

    23Jul, 2024
    What’s self-healing Generative AI? How it works and can revolutionize the finance sector


    Self-healing AI technology is emerging as a groundbreaking development in artificial intelligence, promising to revolutionize how systems manage and resolve issues independently. With the generational capabilities available today, this is becoming a formidable power to bring major changes in the technology landscape. This innovation is particularly transformative in the Banking, Financial Services, and Insurance (BFSI) sector, enhancing reliability, security, and efficiency. But what exactly is self-healing AI, and how does it function?

    Understanding Self-Healing with Generative AI

    Self-healing AI refers to a system that can autonomously detect, diagnose, and repair problems within itself without human intervention. This advanced technology is designed to improve the reliability and efficiency of AI systems by addressing issues as they arise, ensuring continuous and optimal performance.

    “Self-healing technology with Generative capabilities, automates problem resolution, thus significantly reducing the need for manual intervention. This automation liberates IT personnel to concentrate on strategic initiatives such as testing and enhancing their contribution to the organization’s value chain. Moreover, by curtailing downtime and manual troubleshooting, self-healing mechanisms assisted by generative capabilities, offer substantial cost savings and augment operational efficiency,” Rahul Chaudhari, Founder & Chief Executive Officer of Qualitia said.

    How Self-Healing Generative AI Works in BFSI

    Monitoring and Detection: The first step in self-healing AI involves continuous monitoring of the system. Advanced algorithms keep an eye on various parameters and performance metrics to identify any anomalies or deviations from the norm. This real-time surveillance is crucial for the early detection of potential problems in critical BFSI operations such as transaction processing and fraud detection.

    Diagnosis: Once an issue is detected, the system quickly diagnoses the problem. Machine learning models analyze the data to determine the root cause of the anomaly. Based on the systems learning experience, the generative models can generate the right solution. In the BFSI sector, this diagnostic process can be likened to analyzing unusual transaction patterns to identify potential vulnerabilities in software.

    Automated Repair: After diagnosing the issue, the AI system initiates an automated repair process. This could involve restarting certain components, reallocating resources, or applying patches and updates. The goal is to resolve the problem swiftly to minimize downtime and maintain seamless operation, crucial for maintaining customer trust and operational integrity in BFSI services.

    Learning and Adaptation: Self-healing AI systems are designed to learn from each incident. By analyzing the cause and solution of each problem, the AI continuously improves its diagnostic and repair capabilities. This adaptive learning process helps the system become more resilient over time, an essential feature for the BFSI sector where new threats and issues constantly emerge.

    “A paramount advantage of self-healing is the elevation of customer experience. Banks can guarantee the uninterrupted availability of online and mobile banking services, delivering a consistent and dependable experience to customers. Automated detection and resolution of technical issues ensure minimal service disruption, while improved security protocols through self-healing reduce the cost of bad software, safeguarding customer experience,” Rahul Chaudhari from Qualitia said.

    Applications of Self-Healing AI in BFSI

    Self-healing AI has a wide range of applications in the BFSI sector:

    Fraud Detection and Prevention: Self-healing AI continuously monitors transactions for unusual patterns and automatically addresses suspicious activities, thereby enhancing fraud prevention mechanisms.

    Transaction Processing: Self-healing AI ensures that transaction systems remain operational, automatically resolving issues that could delay or disrupt financial transactions.

    Customer Service: In customer service applications, self-healing AI can maintain chatbots and automated support systems, ensuring they function correctly and efficiently without manual intervention.

    “With digitalization at the forefront, BFSI technology teams are tasked with ensuring flawless performance and zero downtime. Today, some products have capabilities to generate automated flows for popular CRM products with just one click. Thereby eliminating the effort of building test automation and making it over ten times faster, with 100% no-code. Hence, the QA leaders can arrest 80% defects right in the QA environment and reduce cost and time of deployment with no backlogs. Such automation not only saves precious time but also empowers these teams to engage in groundbreaking innovation and value creation, directly influencing Return on Investment (ROI) and customer acquisition,” Rahul Chaudhari said.

    “Studies indicate that approximately 75% of automated tests fail due to code alterations, rendering issues, or unhandled behaviors. AI-driven generation addresses this challenge by discerning relevant changes in applications and comparing them against a predefined baseline. Upon recognizing the object in question, the AI adapts it in real-time, allowing the test to proceed without interruption, hence giving a great user experience right from the first release,” Rahul added.

    The Future of Self-Healing AI in BFSI

    As AI technology continues to evolve, the capabilities of self-healing systems are expected to expand. Future advancements may include more sophisticated diagnostic tools, enhanced learning algorithms, and broader applications across different BFSI operations. The ultimate goal is to create AI systems that are not only intelligent but also self-sufficient, capable of maintaining their health and functionality with minimal human intervention.

    Generating AI represents a significant step forward in the quest for more autonomous and reliable artificial intelligence systems in the BFSI sector. By ensuring continuous operation and reducing the need for manual maintenance, this technology holds the potential to transform the industry, making financial services more resilient, secure, and efficient.

    Ref: link

    22Jul, 2024
    Six Opportunities for Scientists and Engineers to Learn Programming Using AI Tools Such as ChatGPT
    Six Opportunities for Scientists and Engineers to Learn Programming Using AI Tools Such as ChatGPT

    In today’s scientific and engineering advancement era, AI technologies are revolutionizing how researchers approach programming and data analysis tasks. Here are some opportunities where AI tools like ChatGPT are making a significant impact: personalized tutorials tailored to specific domains, assistance learning complex data visualization libraries and new programming languages, refactoring and understanding legacy code, and challenging assumptions in scientific coding practices. These tools not only expedite learning and enhance productivity but also encourage deeper critical evaluation, fostering innovation and improving code quality across scientific disciplines.

    OPPORTUNITY 1: Using AI to Create Customized Programming Tutorials for Your Domain

    Harnessing AI tools to create personalized programming tutorials tailored to specific scientific or engineering domains is paramount. Traditional tutorials often use generic datasets such as “iris” or “mtcars,” which may lack relevance to learners outside those fields. For example, a geologist learning Python could benefit from tutorials using geological survey data rather than standard datasets like flower measurements or car specifications.

    OPPORTUNITY 2: Using AI to Learn Complex Data Visualization Libraries

    Scientific programmers frequently navigate intricate data visualization tasks. AI can simplify learning complex libraries like Matplotlib or ggplot2 by generating tutorials that align with specific research needs. For instance, a marine biologist aiming to visualize biodiversity trends could receive tailored tutorials on creating dynamic scatterplots that differentiate fish species by color, shape, and size.

    OPPORTUNITY 3: Refactoring Exploratory Code into Maintainable Software

    In scientific research, initial exploratory code often evolves into long-term solutions requiring clarity and maintainability. AI can assist by analyzing code structures and recommending improvements. For example, AI tools can suggest refactoring options such as modularizing analysis scripts or standardizing variable naming conventions, enhancing code readability and sustainability.

    OPPORTUNITY 4: Understanding Inherited Legacy Code

    Scientists frequently inherit complex legacy codes from previous researchers. AI tools can aid by providing detailed explanations and supplementary documentation to demystify inherited codebases. For example, when inheriting climate modeling algorithms, AI can clarify the logic behind temperature variance calculations, facilitating quicker adaptation and modification for new research purposes.

    OPPORTUNITY 5: Learning New Programming Languages on Demand

    Adapting to new programming languages is crucial in scientific endeavors where specific libraries or tools may necessitate rapid skill acquisition. AI can expedite this process by translating example code snippets between languages. For instance, a bioinformatician needing to implement genetic sequence analysis in Julia can utilize AI to translate existing Python scripts into Julia syntax, accelerating the learning curve within their existing computational workflow.

    OPPORTUNITY 6: Questioning Assumptions in Scientific Code

    Ensuring scientific code aligns with accurate assumptions is critical for reliable research outcomes. AI tools can serve as critical evaluators, prompting researchers to justify algorithmic choices and statistical methods. For example, AI can question the rationale behind statistical tests chosen for genomic data analysis, promoting deeper scrutiny of scientific hypotheses and ensuring robust methodological integrity.

    Conclusion

    AI tools like ChatGPT present significant opportunities for scientists and engineers to enhance their programming proficiency and streamline research workflows. By leveraging AI’s capabilities alongside human expertise, researchers can optimize coding practices, accelerate learning of new technologies, and foster rigorous scientific inquiry.

    Reference: https://ieeecs-media.computer.org/media/marketing/cedge_digital/ce-jul24-final.pdf

    20Jul, 2024
    Recent CrowdStrike outage: What you should know

    On Friday, July 19, 2024, nearly 8.5 million Microsoft devices were affected by a faulty system update, causing a major outage of businesses and services worldwide. This equates to nearly 1% of all Microsoft systems globally and has led to significant disruptions to airlines, police departments, banks, hospitals, emergency call centers and hundreds of thousands of other private and public businesses.

    What caused this outage in Microsoft systems?

    The global outage of specific Microsoft-enabled systems and servers was isolated to a faulty software update released by CrowdStrike, one of the largest independent cybersecurity companies with nearly 30,000 subscribers worldwide.

    With the majority of these subscribers automatically pushing new security updates as they become available, all impacted systems were put into a BSOD (Blue Screen of Death) state. This triggered an infinite boot cycle of the operating system, leaving the systems unable to boot correctly. The operating system then attempts to restart but encounters the same error, causing the process to repeat indefinitely.

    On July 19, CrowdStrike’s CEO George Kurtz announced on X that the company is “actively working with customers impacted by a defect found in a single content update for Windows hosts.” He also confirmed that “This is not a security incident or cyberattack. The issue has been identified, isolated and a fix has been deployed.”

    What has been the aftermath of this incident?

    Even though this CrowdStrike event only affected less than 1% of all Microsoft-enabled systems, the aftermath so far has been significant.

    One of the largest industries impacted by the major outages caused by this faulty update has been air travel. On Friday, more than 3,300 flights had already been canceled around the globe. In the United States, three major airlines Delta, American, and United all grounded their flights for several hours, causing a significant backlog of customer and commercial travel. Airports in Tokyo, Amsterdam, and Delhi were also impacted, creating major issues in other international locations.

    Banking institutions were also significantly disrupted by system outages that impacted everything from ATMs to mobile banking applications and call centers. Even more critical has been the impact on essential emergency services such as hospitals and 911 dispatch teams.

    Massachusetts General Hospital released the following statement regarding the impact of the outage on its operations: “A major worldwide software outage has affected many of our systems at Mass General Brigham, as well as many major businesses across the country. Due to the severity of this issue, all previously scheduled non-urgent surgeries, procedures, and medical visits are canceled today.”

    The ongoing aftermath of this situation has demonstrated how reliant we are on an interconnected ecosystem of technology and services.

    How is this issue being resolved?

    As George Kurtz mentioned in his statement on X, CrowdStrike has already fixed the issue on its end and is actively working with its customers to fully resolve the problem. However, in an interview on the TODAY show on NBC, he also stated, “It could be some time for some systems that just automatically won’t recover.”

    Many IT experts agree with this statement and claim it could take several days for larger organizations to get their systems back to normal operation. The problem lies in the BSOD issues that are being created. This means that while CrowdStrike has “pushed” an automatic update to users on their end since many customers will be unable to fully reboot their system, they won’t be able to receive and install the update.

    CrowdStrike has published manual remediation actions for IT administrators to follow in the event they can’t see an automatic recovery from the issue. These actions involve booting an operating system into “safe mode,” making modifications to the installed drivers, and rebooting again safely.

    While the company has been clear that it has a permanent fix for this issue, it will take time for IT administrators to gain manual access to remote servers and systems running Microsoft operating systems so they’re able to address these issues.

    Unfortunately, the damage has already been done. The ongoing ripple effects of significant economic impacts caused by this event continue to spread, with early estimates of the total losses associated with it nearing $1 billion.

    Ref: https://securityintelligence.com/news/recent-crowdstrike-outage-what-you-should-know/

    19Jul, 2024
    AI vs AI: Using AI to Detect AI-Generated Content

    Artificial Intelligence (AI) has revolutionized the field of content generation, providing tools capable of producing text that is increasingly indistinguishable from that written by humans. These sophisticated algorithms, trained on vast datasets, have mastered the nuances of language, enabling them to generate articles, stories, reports, and more with remarkable proficiency. As AI-generated content becomes more prevalent, the ability to discern between content created by humans and that generated by AI has become a pressing concern. The proliferation of such content has significant implications for various sectors, including journalism, academia, and the broader content creation industry, making the development of detection methods a crucial area of research.

    The Rise of AI in Content Creation

    The integration of AI into content creation marks a transformative moment in the evolution of digital media. Advanced models like GPT-4 have pushed the boundaries, producing high-quality text that can mimic specific writing styles and adapt to diverse content requirements. This capability has led to AI’s adoption across different sectors, from automated news reports to personalized marketing copy. As the technology progresses, the quantity and sophistication of AI-generated content continue to surge, raising important questions about authenticity and trust in digital communication.

    Understanding How AI Generates Content

    AI generates content through machine learning algorithms, particularly those using methods such as deep learning and natural language processing (NLP). By ingesting vast quantities of text, these algorithms learn patterns and structures of language, enabling them to predict and generate coherent and contextually relevant text sequences. The generative process typically involves training a model on a specific corpus of text, after which it can produce new content by sampling from the probability distribution of learned words and phrases. This content can range from simple structured outputs to complex narrative forms, reflecting the intricate nature of the learned linguistic models.

    The Challenge of Detecting AI-Generated Content Using AI

    Detecting AI-generated content is challenging because these systems are designed to replicate human-like text patterns. The subtleties that differentiate AI-written text from human-written text are often minute and can be obfuscated by the AI’s learning from human-tweaked content. As generative AIs become more advanced, the detection process requires increasingly sophisticated techniques, often utilizing similar AI-powered tools to identify nuances and patterns that may indicate machine authorship. The battle to differentiate content is further complicated by the rapid evolution of AI, with each iteration becoming more adept at mimicking authentic human writing styles.

    Methods for Identing AI-Created Texts

    Several methods have been developed to identify AI-created texts, leveraging a variety of linguistic and technical features. Stylometric analysis, for instance, examines writing style, looking for patterns that are atypical of human writing. Machine learning classifiers are trained to distinguish between human and machine writing based on training datasets labeled accordingly. Other methods involve the assessment of semantic coherence, the use of watermarking techniques during AI text generation, or the detection of certain AI-specific artifacts that are left behind in the text. Each of these methods requires constant updating and refinement to keep pace with the evolving capabilities of generative AI models.

    AI-Driven Tools for Content Verification

    To counter the challenges posed by AI-generated text, a new wave of AI-driven tools for content verification has emerged. These tools often utilize the same underlying technologies as the content generators, such as deep learning networks trained to detect anomalies and patterns indicative of AI authorship. Some tools focus on vering the source of the content, while others analyze writing style consistency or unexpected text structures. The key lies in these tools’ ability to adapt and learn from new data, ensuring resilience against the continuously improving quality of AI-generated content.

    The Arms Race: AI Detectors vs. AI Creators

    The dynamic between AI detectors and AI creators is reminiscent of an arms race, where advancements in AI-generated content are met with corresponding developments in detection technologies. As AI creators leverage new techniques to produce more convincing content, AI detectors must evolve, employing deeper and more nuanced analysis to maintain the upper hand. This technological tug-of-war drives innovation in both fields, as each iteration of AI-generated content becomes more sophisticated, so too do the methods and tools designed to detect it.

    The Impact of AI Detection on Digital Media

    The efficacy of AI detection tools has significant ramifications for the integrity of digital media. In an era where information can be rapidly disseminated and consumed, the ability to verify the authenticity of content is paramount. Reliable detection methods are essential to maintain trust in digital platforms, safeguard against misinformation, and protect intellectual property rights. The media industry, in particular, relies on these tools to uphold journalistic standards and ensure the credibility of published content.

    Ethical Considerations in AI Content Detection Using AI

    Ethical considerations in AI content detection revolve around privacy, bias, and the potential for misuse. Detection tools must navigate the fine line between scrutiny and invasion of privacy, ensuring that legitimate content is not unfairly targeted. Additionally, there is a risk of bias in detection algorithms, which must be addressed to prevent discrimination against certain types of content or authors. Lastly, there is the danger of these tools being used to suppress or manipulate information. As such, transparency in the functioning of AI detection systems is critical to ensure they are used responsibly and do not become tools for censorship.

    The Future of AI in Content Authenticity

    Looking forward, the interplay between AI-generated content and AI-driven authenticity checks is set to become even more intricate. As AI continues to advance, we may see the emergence of new standards and regulatory frameworks guiding the use of AI in content creation and verification. The development of universally accepted benchmarks for AI transparency, such as content origin certificates or the equivalent of ‘nutrition labels’ for information, could play a pivotal role in managing the impact of AI-generated content. Moreover, ongoing research is likely to yield more robust detection mechanisms that can keep pace with AI’s capabilities, ultimately contributing to a more trustworthy digital ecosystem.

    Conclusion: AI vs. AI

    The advent of AI-generated content challenges our traditional understanding of creativity and authorship. Yet, as AI detection methods become more sophisticated, there is potential for a symbiotic relationship between human and artificial creativity. Rather than viewing AI as a threat to human content creators, it can be seen as a tool that complements human ingenuity, with detection technologies ensuring the integrity of the content. The key lies in striking a balance that leverages the strengths of AI to enhance human creativity while maintaining transparency and trust in the content that shapes our world.

    Ref: Link

    Akchara