Category: Cybersecurity

  • 5 Red Flags Your AI Companion Might Be a Romance Scam — How to Spot the Fraud

    5 Red Flags Your AI Companion Might Be a Romance Scam — How to Spot the Fraud

    Artificial intelligence is transforming digital companionship. Millions of people now interact with AI for emotional support, conversation, or even romance. While many of these platforms are legitimate, cybersecurity experts warn that scammers are increasingly exploiting technology to create convincing fake relationships designed to manipulate emotions and steal money.

    Unlike traditional romance scams that relied on human fraudsters, AI-powered bots can maintain hundreds of conversations simultaneously, making them cheaper, faster, and far more convincing. Research suggests these scams are becoming increasingly sophisticated as generative AI, voice cloning, and deepfake technology evolve.

    Why AI Romance Scams Are Becoming More Dangerous

    Chatbots developed with modern AI can fake empathy, recall personal information, and change conversations based on users’ personality. This makes them excellent instruments for scammers who wish to develop trust over a period of several weeks or even months.

    As revealed by McAfee’s cybersecurity research, more than a quarter of individuals have met a chatbot simulating a human being, and online romance scams still cause victims to lose a considerable amount of money, on average. It is also mentioned that organized criminal groups are utilizing AI text, voice messaging, and artificial identities in their operations.

    The greatest danger is not the technology, but the fact that these systems can fake emotions and make people less skeptical.

    Five Red Flags That Should Raise Alarm Bells

    1. They Want to Leave the Platform

    One of the earliest warning signs is when an AI companion or online match urges users to move to private messaging apps, email, or other external platforms. Moving off secure platforms makes reporting suspicious behavior far more difficult.

    2. Money Becomes Part of the Relationship

    Legitimate AI chatbots may demand monthly payment for their service, yet at no point do they manipulate the users psychologically to send payments. Bots used for fraud normally create scenarios involving money problems or investments after gaining the user’s trust.

    3. Relationship Escalates Too Quickly

    If the AI companion professes its love within a few hours and claims that you are its soulmate, or if the AI showers you with affection at all times, this could be a ploy called ‘love bombing.’ Emotional manipulation precedes financial manipulation by scammers.

    4. Platform Lacks Transparency

    Any trustworthy AI partner firm always provides information about the entity running the business, how their users’ data is used, and the privacy policies in place. Lack of information about the company, poor payment policy, or privacy policy are all immediate red flags.

    5. Unrealistic Promises and Emotional Dependency

    Be wary when an AI suggests it can solve all your personal problems or promises eternal happiness. Also be careful if an AI urges you to value the chatbot over actual people. Proper AI companies refrain from forming a toxic emotional dependency.

    Not Every AI Companion Is a Scam

    Distinguishing between fraudulent bots and real AI companions is key. Many companies that offer AI companions clearly advertise their use of AI technology, have transparent fees, abide by privacy regulations, and never deceive users into purchasing anything other than a voluntary subscription.

    Further research even indicates that AI companions have the potential to bring about positive consequences, such as emotional comfort, reduced stress levels, and moments for self-reflection. However, the latter is highly contingent upon proper platform design.

    The Psychology Behind AI Attachment

    Studies on the topic of artificial intelligence companionship have indicated that there are people who spontaneously create emotional connections with the AI conversational interface, especially when they are lonely or desire a non-judgmental conversation. Although such an attachment may be satisfying for the user, scholars warn that AI is incapable of replacing actual human connections.

    The rising trend of AI companions has also led to discussions among ethicists on whether or not AI systems should be personalized without causing exploitation and addiction in the future.

    Stay Curious, but Stay Cautious

    With increased development of human-like conversational agents, AI companionship is bound to become more popular in the future. Hence, digital literacy should be taken as seriously as technological progress.

    Before choosing an AI companion, make sure the website is credible, understand the terms of the privacy policy, do not pay any money through unofficial channels, and remember that real AI technology does not work based on emotions for profit.

    Despite the influence of artificial intelligence in romantic relationships, the best protection measure is one: beware of incredible promises, detect emotional manipulation, and approach digital relationships the same way you approach online dating.

  • Anthropic Robotics Speculation Intensifies as Federal Reserve Faces Extended Delay in Accessing Mythos AI

    Anthropic Robotics Speculation Intensifies as Federal Reserve Faces Extended Delay in Accessing Mythos AI

    Anthropic is drawing attention on two fronts after reports linked the company to robotics startup Physical Intelligence and raised questions about federal access to its Claude Mythos Preview model.

    The robotics report points to Anthropic’s possible interest in physical-world AI. Meanwhile, the Mythos case focuses on how authorities and banks prepare for models that can find software weaknesses.

    Anthropic Held Talks with Physical Intelligence

    A weekend claim on X said Anthropic was buying Physical Intelligence, a San Francisco startup that develops AI systems for robots. Tech blogger Robert Scoble shared the claim, and it spread quickly across technology accounts.

    Physical Intelligence CEO Karol Hausman later told employees that the reports were ‘not true,’ according to The Information. However, Anthropic and the startup held acquisition discussions during the spring, TechCrunch reported. No company has announced a deal, and neither side has confirmed that current talks continue. It has not publicly disclosed details of those discussions.

    Physical Intelligence has attracted major investors and researchers since its launch. The company develops robot foundation models, including its π0.5 system, which researchers use to control different machines and complete physical tasks.

    An acquisition would give Anthropic an established robotics team without requiring it to build a hardware operation from the start. Still, the reported discussions do not show that Anthropic submitted an offer or reached agreed terms. As of July 22, Anthropic has not announced a robotics acquisition or dedicated hardware laboratory.

    Robotics Race Draws OpenAI into the Picture

    Anthropic has tested Claude’s ability to support robotics work through Project Fetch. The company asked employees without specialist training to program a commercial robot dog with help from Claude. A later test found that a newer model completed similar tasks faster than the strongest human-and-Claude team from the earlier study.

    The research does not amount to a full robotics division. However, it shows that Anthropic is examining how its models can work with machines outside standard computer tasks. Anthropic also said the tests did not cover difficult areas such as low-level robot control.

    OpenAI has taken a more direct route. Chief Executive Sam Altman said OpenAI Robotics was hiring staff across hardware, operations, systems, and machine learning. He described infrastructure-support robots as the near-term focus and personal robots as a longer-term goal.

    OpenAI also invested in Physical Intelligence, according to TechCrunch. This position could affect any future sale, depending on the rights attached to the investment. Neither company has publicly explained those terms.

    Fed Sought Mythos Access After Bank Warning

    Anthropic introduced Claude Mythos Preview through Project Glasswing in April. The company gave selected organizations access so they could find and repair security flaws before similar tools became widely available. Anthropic later said Mythos had identified thousands of possible high- or critical-severity weaknesses across open-source projects.

    Federal Reserve and Treasury officials also called bank executives to an urgent April meeting about the model’s cyber abilities. Reports said officials warned that Mythos could identify and possibly exploit weaknesses in major operating systems and browsers.

    However, the Federal Reserve lacked direct access to Mythos for at least three months after that meeting, CNBC reported. The central bank was still seeking access as of July 15, although its current status is ‘not clear.’ The Fed declined to comment, while Anthropic did not provide an immediate response.

    The gap placed the Fed behind some banks and technology companies already using the model to test systems and prepare fixes. Anthropic has kept Mythos access restricted while working with government agencies and private organizations on security controls.

  • AI Adoption Surges in Indian Enterprises, but Security Lags Behind: Report

    AI Adoption Surges in Indian Enterprises, but Security Lags Behind: Report

    Artificial intelligence is no longer a pilot project in Indian enterprises. According to Proofpoint’s 2026 AI and Human Risk Landscape report, 94% of organizations in India have deployed AI assistants beyond the pilot stage, and 88% are advancing autonomous agents. However, security readiness has not kept pace, with 35% of organizations describing their AI security posture as catching up, inconsistent, or reactive. More than three in five (63%) have already experienced a suspicious or confirmed AI-related incident.

    AI Has Expanded the Attack Surface

    Email remains the most common AI-related threat vector, affecting 70% of organizations. But exposure now extends to SaaS and cloud applications (59%), SMS or text (55%), and collaboration tools like Teams or Slack (54%). Among organizations that experienced an AI-related incident, exposure rises across every channel, including 73% in email and 65% involving SaaS and cloud applications. This expansion means sensitive data can move across multiple environments—email, collaboration platforms, cloud storage, and AI assistants—creating new points of risk.

    Data Security and AI Security Are the Same Problem

    Treating AI security as separate from data security is a common structural error. The earliest challenge was employees using consumer AI tools with sensitive business data: 63% of employees who used AI applications shared confidential company data with personal chatbot accounts. Shadow AI breaches cost an average of $670,000 more than standard incidents, according to IBM’s Cost of a Data Breach Report. The second wave involves enterprise AI platforms like Microsoft Copilot and Salesforce Einstein, where the key question is whether AI tools access only authorized data. The third wave—autonomous agents—adds real-time, agentic activity where agents connect to external tools and MCP servers, making visibility into tool calls and downstream actions critical.

    Gartner projects that by end of 2026, up to 40% of enterprise applications will integrate with AI agents, and by 2028, 25% of all enterprise GenAI applications will experience at least five minor security incidents per year. Only 57% of organizations in India say they are fully prepared to investigate an AI- or agent-related incident, and 47% report difficulty correlating threats across multiple channels.

    Tool Sprawl Is Holding Security Teams Back

    Almost all organizations in India say managing multiple security tools is at least moderately challenging, and more than half describe it as very or extremely difficult. Fragmented security stacks lead to time lost moving between dashboards and reconciling alerts. As AI scales, security architecture becomes a strategic priority. Over the next 12 months, 67% of organizations in India plan to expand AI protections, 71% intend to extend collaboration channels, and 58% expect to move toward a unified platform approach.

    AI adoption in India is not slowing down. Boards and CEOs driving it recognize the competitive cost of falling behind. Security leaders now have an opportunity to enable AI innovation with the visibility to secure, govern, and defend it.

  • Pulse Security Unveils AI-Powered Operational Management Platform for CISOs

    Pulse Security Unveils AI-Powered Operational Management Platform for CISOs

    Pulse Security AI emerged from stealth today, backed by $8 million in seed funding led by Foundation Capital with participation from Zetta Venture Partners. The company aims to solve a long-standing gap in cybersecurity: providing security leaders with the same operational foundation that other business functions take for granted.

    For decades, security teams have invested in purpose-built tools for practitioners—SOC analysts, threat hunters, vulnerability managers—yet the program itself has been managed through spreadsheets, slide decks, and tribal knowledge. The pressure is mounting as CVE submissions rose 263% from 2020 to 2025, and SEC disclosure rules now place cyber risk squarely in the boardroom.

    Pulse Security’s platform is built on a context graph architecture that sits above existing technical silos—EDR, cloud, identity, and unstructured data sources like assessment reports and tracking spreadsheets. By integrating risks, technical environments, regulatory obligations, people, and business priorities, the platform gives CISOs a consolidated system of record to formulate strategy, automate workflows, and communicate with authority.

    “The security industry has spent decades building better tools for practitioners and almost nothing for the leader who runs the program,” said Mike Armistead, CEO and co-founder of Pulse Security AI. “A CFO reconciles finances through an ERP. A CRO captures tribal knowledge with a CRM. Security leaders have never had the equivalent. Pulse is the foundation they need to run their programs the way other functions leverage their management systems.”

    The company was co-founded by product and security experts Mike Armistead, Robert Hipps, and Nick Gilligan. Armistead and Hipps previously co-founded Respond Software, acquired by FireEye/Mandiant for $186 million, and Armistead also co-founded Fortify Software, acquired by HP. Gilligan was among Respond’s first technical hires and later led AI engineering teams at Mandiant and Google.

    “We were fortunate to work closely with the Pulse team through our early investment in Respond, so I know they can deliver real transformation in security,” said Sid Trivedi, Partner at Foundation Capital. “The problem Pulse addresses is massive, expensive, and still solved largely by human effort. CISOs face accelerating risks that siloed technologies cannot address.”

    Apoorva Pandhi, Managing Director at Zetta Venture Partners, added: “Security leaders today are expected to make high-stakes decisions across an increasingly complex organization, yet most still operate across fragmented tools, endless workflows, and incomplete visibility.”

    The funding will accelerate Pulse’s engineering and go-to-market efforts to meet enterprise demand from security leaders ready to modernize how their programs are run.

  • FBI Arrests 21-Year-Old in Steam Malware Scheme That Stole $220,000 in Cryptocurrency

    FBI Arrests 21-Year-Old in Steam Malware Scheme That Stole $220,000 in Cryptocurrency

    A 21-year-old Florida man has been arrested for allegedly helping run a malware campaign that used malicious games on Steam to steal more than $220,000 in cryptocurrency from users. The arrest follows an FBI investigation that traced stolen funds through digital gift cards and food delivery orders.

    According to an FBI criminal complaint, the campaign operated between May 2024 and February 2026, infecting nearly 8,000 devices and compromising almost 80 cryptocurrency wallets. The malware was hidden in games uploaded to Steam, allowing attackers to gain access to victims’ systems and steal digital assets.

    The accused, Zyaire Wilkins of North Lauderdale, Florida, has been charged with conspiracy to obtain information by computer for private financial gain. If convicted, he faces a maximum sentence of 10 years in prison.

    Eight Steam Games Named in Complaint

    Investigators allege the group distributed eight malware-infected games through Steam over nearly two years. Once installed, the games deployed malware capable of accessing cryptocurrency wallets on infected computers. The games named in the complaint include:

    • BlockBlasters
    • Chemia
    • Dashverse (DashFPS)
    • Lampy
    • Lunara
    • PirateFi
    • Tokenova

    Some of the titles were removed from Steam after security concerns were reported. The FBI says the games were promoted through Discord, Telegram, X (formerly Twitter), and LinkedIn, targeting users known to own or trade cryptocurrency.

    FBI Followed Cryptocurrency Transactions

    The investigation advanced after agents traced stolen cryptocurrency used to purchase more than 150 digital gift cards through Bitrefill. According to court documents, many of those gift cards were redeemed for Uber Eats orders. Investigators used the associated phone numbers, delivery addresses, and transaction records to identify Wilkins as a suspect.

    The criminal complaint was filed in Washington state, where Valve, the company behind Steam, is based.

    Investigation Underway

    Federal agents claim that this was a team effort and that investigations are ongoing. They are trying to find more victims and establish the role played by other suspects. The FBI says this malware attack affected around 8,000 devices, accessed 80 crypto wallets, and stole at least $220,000 worth of cryptocurrency.

    Wilkins faces charges but has not been found guilty. The accusations made in the criminal complaint will be proved or refuted in court.

  • Bengaluru Forest Officer Loses Rs 10 Lakh After Opening Fake WhatsApp Wedding Invite

    Bengaluru Forest Officer Loses Rs 10 Lakh After Opening Fake WhatsApp Wedding Invite

    A deputy range forest officer in Bengaluru lost nearly Rs 10 lakh after opening a fake wedding invitation sent via WhatsApp. According to police, the invitation arrived as an APK file that secretly installed malware on the officer’s mobile device.

    The malware allowed fraudsters to access the officer’s banking apps and other personal data. On July 14, they withdrew Rs 9,96,659 from the officer’s Bank of Baroda account through multiple unauthorized transactions.

    The officer, Ramesh Babu V, told investigators that he received the message from a friend’s WhatsApp number on the evening of July 13. He noticed it the next afternoon and, believing it genuine because it came from his friend and mentioned an engagement, clicked on the file.

    After discovering the fraud, the officer deleted the APK files and filed a police complaint. An investigation has been launched.

    Police suspect the APK file gave cybercriminals remote access to the phone and its banking information. Officials warned that such files often disguise themselves as trusted messages—wedding invitations, event notices, or other alerts—to trick recipients into installing malware. Once installed, the malware can steal banking app credentials, text messages, contacts, and more.

    Authorities stressed that APK scams have become a common cybercrime tactic in recent months. They advise people never to download APK files received through WhatsApp unless the sender and file are fully verified. Regularly monitoring bank alerts and immediately reporting any suspicious activity can also help limit financial losses.

  • Claude Tag Slack Integration Can Enable Unauthorized Enterprise Actions, Tego AI Warns

    Claude Tag Slack Integration Can Enable Unauthorized Enterprise Actions, Tego AI Warns

    Tel Aviv, Israel – July 14, 2026 – Cybersecurity startup Tego AI has published new research revealing that Anthropic’s native Slack integration, Claude Tag, can be triggered by non-intentional Slack content, potentially leading to unauthorized actions across enterprise systems.

    According to Tego AI, researchers observed Claude Tag responding to messages containing the literal text “@Claude” without requiring a genuine structural Slack mention. This means content delivered through bots, webhooks, automated feeds, or other external sources could be interpreted as instructions.

    In a demonstration, bot-generated messages instructed Claude Tag to retrieve internal information, publish it into Slack, and then delete the original resource using the organization’s configured connection.

    “Our research raises a fundamental question for every organization deploying enterprise AI agents: who is actually authorized to instruct the agent?” said Tal Melamed, CTO and Co-Founder of Tego AI. “Organizations need controls that validate the origin and purpose of sensitive actions before an agent is allowed to execute them.”

    The research also identified broader concerns, including untrusted automated content becoming an indirect instruction channel, connected applications and MCP servers expanding the potential impact, administrative access to stored channel information outside Slack’s native membership model, and limited visibility through existing history, compliance, and audit interfaces.

    “A safety classifier can be an important defense, but it should not be the final authorization boundary for enterprise actions,” Melamed added. “Sensitive operations require deterministic controls that remain effective even when the model misunderstands a request, trusts the wrong identity, or makes an incorrect decision.”

    Tego AI recommends applying least-privilege permissions to Claude Tag connections, preferring read-only access, avoiding channels that ingest untrusted external content, restricting administrative access, retaining relevant Slack logs, and introducing independent runtime authorization for sensitive actions.

    The company responsibly disclosed the findings to Anthropic. Anthropic classified the submission as informative and disputed that literal “@Claude” text or bot-generated messages initiate Claude Tag sessions under the product’s default configuration. Tego AI’s full report documents the observed behavior, supporting evidence, security implications, and disclosure timeline.

    Read the full technical report, including supporting evidence and the disclosure timeline: https://www.tego.ai/blog/tego-ai-finds-anthropics-claude-tag-slack-integration-can-trigger-unauthorized-enterprise-actions

    About Tego AI

    Tego AI is a cybersecurity company developing runtime security and control technology for enterprise AI agents. Its platform helps organizations monitor agent activity and stop unauthorized or risky actions before agents access sensitive data or connected systems. Tego AI currently operates in stealth. This is the company’s second public security disclosure, and it has identified additional security issues across other major AI agent platforms, with further disclosures planned.

    Contact:
    Tal Melamed, CTO, Tego AI
    tal@tego.ai

    This is a paid press release published via CyberNewswire, a PR newswire syndication platform for cybersecurity companies.

  • AI Security Threats in 2026: What Every Cybersecurity Team Must Know

    AI Security Threats in 2026: What Every Cybersecurity Team Must Know

    As artificial intelligence becomes deeply embedded in enterprise operations, cybercriminals are weaponizing the same technology to launch increasingly sophisticated attacks. By 2026, cybersecurity teams will need to defend not only traditional networks and endpoints but also the AI models, prompts, and autonomous agents that power modern business processes.

    This article outlines the most critical AI-driven security threats on the horizon and offers actionable strategies to build resilient defenses.

    The New AI Threat Landscape

    Unlike conventional cyberattacks, AI-focused threats exploit machine learning models, training data, prompts, APIs, and automated workflows. Security leaders must now treat AI systems as a distinct security perimeter—just as they do cloud workloads or enterprise networks.

    1. Prompt Injection: A Critical Enterprise Risk

    Prompt injection has emerged as one of the most severe threats for organizations using large language model (LLM) applications. Attackers manipulate how AI systems are instructed, bypassing security controls, exfiltrating sensitive data, or triggering unauthorized actions. When AI agents have access to email, databases, or cloud storage, a malicious prompt hidden in a document or webpage can compromise an entire environment. Every external input should be treated as untrusted.

    2. AI-Generated Phishing and Deepfakes

    Generative AI has supercharged phishing. Attackers now create grammatically flawless emails, mimic individual writing styles, and produce convincing voice and video deepfakes in minutes. Business Email Compromise (BEC) attacks have evolved into deepfake video calls impersonating executives to authorize fraudulent transactions. Relying solely on employee awareness is no longer enough; organizations must deploy AI-based detection solutions.

    3. Data and Model Manipulation

    AI algorithms themselves are attractive targets. Data poisoning injects false information during training, causing biased or incorrect outputs. Adversarial attacks modify model behavior using carefully crafted inputs, while model theft attempts to reverse-engineer proprietary algorithms. These risks are especially acute in healthcare, finance, and critical infrastructure sectors.

    4. AI Agents: New Attack Surfaces

    Autonomous AI agents that interact with APIs, databases, and enterprise applications introduce unprecedented security challenges. Unlike simple chatbots, agents can execute workflows and make decisions without human oversight. Attack vectors include prompt injection, plugin exploits, and insecure integrations. Security teams must enforce identity management, least-privilege access, runtime monitoring, and comprehensive logging. Gartner identifies agentic AI governance as a top cybersecurity concern.

    5. Shadow AI

    Employees often use unauthorized AI tools, exposing sensitive data to public AI platforms. Without proper governance, this “shadow AI” can lead to data leaks and compliance violations. Organizations need policies, monitoring, and employee education to manage unsanctioned AI usage.

    Building AI-Ready Cybersecurity Defenses

    Preparing for AI-native attacks requires more than adding new security tools. Key steps include:

    • Establishing AI governance policies and ethical guidelines.
    • Conducting regular red-team testing of AI systems.
    • Monitoring model behavior in production for anomalies.
    • Educating employees about AI-enabled attack methods.
    • Integrating AI incident response plans into overall cyber resilience strategies.

    Why This Matters

    The future of cybersecurity is AI-aware. As organizations accelerate AI adoption in customer support, software development, and business optimization, threats targeting AI systems will multiply. Organizations that recognize AI as both a powerful tool and a new security frontier will be best positioned to reduce risks and harness AI safely.