<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by zkHive on Medium]]></title>
        <description><![CDATA[Stories by zkHive on Medium]]></description>
        <link>https://medium.com/@zkhive?source=rss-21b4d1e86838------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*-WR2uB9rKofnzG-_Y1pxFg.jpeg</url>
            <title>Stories by zkHive on Medium</title>
            <link>https://medium.com/@zkhive?source=rss-21b4d1e86838------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 01:53:17 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@zkhive/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[HiveBot For Groups Is Now Live!]]></title>
            <link>https://medium.com/@zkhive/hivebot-for-groups-is-now-live-20dc1b3888d9?source=rss-21b4d1e86838------2</link>
            <guid isPermaLink="false">https://medium.com/p/20dc1b3888d9</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[ethereum]]></category>
            <category><![CDATA[zero-knowledge-proofs]]></category>
            <category><![CDATA[blockchain]]></category>
            <dc:creator><![CDATA[zkHive]]></dc:creator>
            <pubDate>Tue, 26 Mar 2024 16:58:35 GMT</pubDate>
            <atom:updated>2024-03-26T16:58:35.063Z</atom:updated>
            <content:encoded><![CDATA[<h3>HiveBot Anti-Virus For Groups Is Now Live!</h3><h4>Every PC has an Anti Virus, and now every group has HiveBot Anti-Virus!</h4><p>Less than 2 weeks after launch, HiveBot’s — zkHive’s first decentralized firewall for Telegram — has already scanned more than 10,000 files for hundreds of users, detected nearly 100 malicious files &amp; links and saved millions of $ in potential losses!</p><p>Up until now, HiveBot was only accessible to individuals — now it is time to go one step further and open HiveBot to every single group!</p><h3>Protect Communities with AI</h3><p>Telegram has become the de-facto city square for crypto enthusiasts, containing more than 70% of crypto-related launches and discussions.</p><p>However, TG communities security state is at an all-time low, with <strong>thousands of hacks and exploits per month</strong>, resulting in losses of millions of dollars <strong>per year</strong>.</p><p>Every TG group contains phishing attacks, fake links, fake airdrops, drainers, malwares and malicious files — without any protection.</p><p>Now all of it is about to change — thank’s to ZKHIVE’s decentralized antivirus — The HiveBot!</p><h3>Anti Virus for every TG group</h3><p>Powered by connectivity AI, zkHive is a state-of-the-art security product that constantly filters out potentially malicious activity in order to keep crypto communities free of fraudulent content, bots and malicious actors before they get a chance to act.</p><p>You can now scan any file or url, directly in every community that is protected by the HiveBot!</p><p><strong>Forget about downloading</strong> suspicious files, links, PDFs, voice messages or any other potentially malicious content (PMC) that someone sends in a group you are part of, simply quote it with the <em>/scan</em> command and HiveBot will immediately perform a full analysis of the file or link and reply with the result.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EZDl7R9rXSBckfNqgbQnYg.png" /><figcaption>Scan every file</figcaption></figure><p>The bot works by uploading your PMC sample and to our secure sandbox in the cloud and then perform multiple checks and tests to check if the given file is safe or not.</p><p>These tests include an aggregation of the most advanced security tools, data sources, online databases and our own proprietary algorithms, which is exactly what makes HiveBot the most advanced product of its kind!</p><p>For a full list of the scan steps, please check out our <a href="https://medium.com/@zkhive/introducing-zkhives-individual-protection-suite-22ef301f159b">previous article</a></p><h3>How To Use The HiveBot In Your Group</h3><p>Using HiveBot in groups is as easy as it gets — simply quote the potentially malicious link, file, pdf or voice message and send the <em>/scan</em> command in the chat, and HiveBot will scan it and send the results to the group chat for everyone to see!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JkKnOfcCg-kh0o8KAEp2Ug.png" /><figcaption>Easy to scan</figcaption></figure><h4>Steps:</h4><ol><li>Add <a href="https://t.me/zkHiveBot">@ zkHiveBot</a> to your groups</li><li>Make it an admin</li><li>Now everyone can start using the bot!</li><li>To scan a file, just quote reply it with /scan reply</li><li>To scan a url: <br>— Quote reply the message that contains the url with /scan <br>— Use the <em>/scan </em>command like this: <em>/scan url_address</em></li><li>to report a file / url, simply quote reply it with <em>/report</em></li></ol><h3>Next-Level Revenue Generation</h3><p>We are happy to announce that HiveBot will be available <strong>for free to all groups for a limited time!</strong></p><p>After this promotional period, HiveBot will incorporate a freemium plan that includes higher rate limits and premium paid features!<br>In addition, HiveBot will also incorporate an ad system, leveraging the massive exposure our platform holds!</p><p>A large portion of the revenue generated by the ecosystem from the aforementioned methods as well as the ones we described in our <a href="https://medium.com/@zkhive/zkhive-the-revenue-model-87296059664b">revenue model article</a>, will be used to buy-back and burn $ZKHIVE from the market, making it deflationary and keeping a constant buy pressure!</p><p>HiveBot for groups marks the first major step in $ZKHIVE’s business plan unleashing the true potential of this 1B$ opportunity!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=20dc1b3888d9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI URL Scanning Is Now Live On HiveBot]]></title>
            <link>https://medium.com/@zkhive/ai-url-scanning-is-now-live-on-hivebot-bf1a7acf46f8?source=rss-21b4d1e86838------2</link>
            <guid isPermaLink="false">https://medium.com/p/bf1a7acf46f8</guid>
            <dc:creator><![CDATA[zkHive]]></dc:creator>
            <pubDate>Tue, 19 Mar 2024 14:29:38 GMT</pubDate>
            <atom:updated>2024-03-19T14:29:38.538Z</atom:updated>
            <content:encoded><![CDATA[<h4>Seamlessly scan any URL and Domain through our decentralized AI HiveBot</h4><p>The Second half of zkHive’s Individual protection suite that was <a href="https://medium.com/@zkhive/introducing-zkhives-individual-protection-suite-22ef301f159b">previously introduced</a> on our blog is going live !</p><p>The most advanced real-time AI security scan for links and domains is now here, making sure no one falls victim to malicious and phishing websites.</p><h3>How It Works</h3><p>Just like HiveBot’s real-time file scanning feature that allows you to scan every file by simply forwarding it to the bot BEFORE you download it, scanning a link is as easy as sending it to HiveBot BEFORE you even click it!</p><p>HiveBot automatically detects any link sent to it and performs a real-time analysis of the website and domain using some of the most advanced public and proprietary security tools available today.</p><p>These tests include an aggregation of the most advanced security tools, data sources, online databases and our own proprietary algorithms!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/796/1*d1WRon2O4fxBwtrVvOUI1A.png" /></figure><h3>Security Features</h3><p>As with file scans, every link goes through a thorough analysis process that includes the same list of security features and checks:</p><p><strong>AntiVirus scans, Anti Malware scans, online database lookup, WebRisk analysis, aggregated user reports and finally our own proprietary HiveMind analysis.</strong></p><p>These steps ensure that HiveBot is the single most up to date security hub right at your fingertips!</p><h3>Bees Watch Each Other’s Back</h3><p>As with any security solution, it may take hours or even days to detect a new attack campaign, just like the Weaponized PDF attack in LATAM we <a href="https://medium.com/@zkhive/zhkhive-saved-millions-from-being-drained-in-the-last-72-hours-956b11fc4b6a">covered previously</a>.</p><p>This emphasizes the importance of HiveBot’s users and our powerful report mechanism!<br>If you see any malicious website — report it using the report button of HiveBot to let other users know to it might not be safe!</p><blockquote><em>“Just like the working bees in the hive work together to keep each other safe, zkHive combines the knowledge of thousands of groups and hundreds of thousands of users to keep the entire network safe” — zkHive CEO</em></blockquote><h3>Reporting And Scanning URLs</h3><p>Reporting URLs has never been easier.<br>Just go to @ ZKHiveBot , and type <em>/report the_phishing_url .<br></em>For example: /report virus.com</p><p>To scan a new URL, simply go to the HiveBot at @ zkHiveBot and forward the URL.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bf1a7acf46f8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[zhKhive Saved Millions From Being Drained In The Last 72 Hours]]></title>
            <link>https://medium.com/@zkhive/zhkhive-saved-millions-from-being-drained-in-the-last-72-hours-956b11fc4b6a?source=rss-21b4d1e86838------2</link>
            <guid isPermaLink="false">https://medium.com/p/956b11fc4b6a</guid>
            <dc:creator><![CDATA[zkHive]]></dc:creator>
            <pubDate>Mon, 18 Mar 2024 10:57:09 GMT</pubDate>
            <atom:updated>2024-03-18T10:57:09.040Z</atom:updated>
            <content:encoded><![CDATA[<h4>A malicious PDF rat is draining millions of dollars, and HiveBot is the only one that can save you</h4><p>In the last 72 hours, the HiveBot saved 7 users from being completely drained by a new a type of a malicious PDF.</p><p>The recent attack campaign that was revealed this week targeted individuals and crypto funds, and could potentially steal tens of millions of dollars.</p><p>This attacks and many similar to it could be mitigated or even <strong>completely</strong> <strong>halted using zkHive’s existing personal security solution — The HiveBot.</strong></p><h3>The New Weaponized PDF Rat Attack</h3><p>An ongoing campaign targeting Latin America over the past few weeks has been uncovered by Any.run.</p><p>The attackers lured victims into downloading a malicious PDF that would eventually install a RAT (Remote Access Tool) on their machine.</p><p>More than 50 samples associated with this campaign have been detected so far.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4IY0uxy1b9SPt6mkCXKjNw.png" /><figcaption>Attack Campaign flow. Source: Any.run</figcaption></figure><p>As cybersecuritynews.com reports: “Cybersecurity experts warn that while this campaign focuses on Latin America, similar tactics could be employed against targets in other regions.<br>Organizations and individuals must remain vigilant, educate themselves on these threats, and employ robust security measures to protect against such sophisticated attacks.”</p><h3>Targeting Crypto Traders On Telegram</h3><p>In this recent attack, the hackers are pulling off a sneaky move targeting influencers, VC owners and community traders.</p><p><strong>Here’s how it went down</strong>: The hackers disguised the harmful file as a White Paper and sent it out to VC owners and influencers. <br>People unsuspectingly opened it, thinking they were getting valuable insights.</p><p>Little did they know, that innocent-looking document was actually packed with malware, ready to wreak havoc on their systems and steal sensitive information — Including your wallets.</p><p>An innocent PDF while disguised as a white paper, could actually wipe out all of your wallets, including an access to your social media and sensetive information.</p><h3>zkHive To The Rescue</h3><p>Using HiveBot, zkHive’s innovative personal firewall on Telegram, this attack could have been swiftly detected and mitigated.</p><p>All the users have to do is toupload the suspicious PDF, just like any other file or link they are sent from unknown sources, and our security system will instantly report this file as potentially malicious and show dozens of other users that fell victim to this attack!</p><p>You do not need to download any file! Simply send it to @ zkHiveBot — and we will do the rest, in a fast and seamless way!</p><p><strong>This is another proof that HiveBot is solving a real issue, potentially saving millions of $ from this campaign alone.</strong></p><p>We once again encourage all users to adopt better security practices, including using HiveBot for scanning every PDF, file, link or voice message and actively report anything suspicious on our platform to warn other users</p><h3>Using HiveBot</h3><p>Using HiveBot has never been easier! <br>Just go to our bot on TG @ zkHiveBot , forward your file without even downloading it — and the HiveBot will do the rest!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=956b11fc4b6a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Introducing: zkHive’s Individual Protection Suite]]></title>
            <link>https://medium.com/@zkhive/introducing-zkhives-individual-protection-suite-22ef301f159b?source=rss-21b4d1e86838------2</link>
            <guid isPermaLink="false">https://medium.com/p/22ef301f159b</guid>
            <category><![CDATA[security-services]]></category>
            <category><![CDATA[eth]]></category>
            <category><![CDATA[zero-knowledge]]></category>
            <dc:creator><![CDATA[zkHive]]></dc:creator>
            <pubDate>Thu, 14 Mar 2024 16:48:00 GMT</pubDate>
            <atom:updated>2024-03-14T16:48:00.228Z</atom:updated>
            <content:encoded><![CDATA[<h4>The first decentralized protection bot is now LIVE on Telegram!</h4><p>Right on time for Pie-day, we are happy to introduce our first product release!</p><p>Forget about hacks, malicious voice messages, files, ZIPs, RARs, links and PDFs!<br>Powered by the most advanced security sources available today, zkHive’s Individual Protection Suite is your own personal firewall, that allows you to <strong>scan every file without downloading</strong> it, directly on Telegram!</p><h3>How It Works</h3><p><strong>Before even downloading</strong> a suspicious file, link, PDF, voice message or any other potentially malicious content (PMC) that someone sends you, simply forward to the HiveBot will activate the <strong>Safe Messages</strong> feature.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pYvOLGfwhtP5wLTUM-OD_w.png" /><figcaption>Just forward the message to @ zkhiveBot</figcaption></figure><p>This feature performs real-time analysis of every file, link, voice message or other potentially malicious file sent to the bot using the most advanced security solutions available today.</p><p>The bot works by uploading your PMC sample and to our secure sandbox in the cloud and then perform multiple checks and tests to check if the given file is safe or not.</p><p>These tests include an aggregation of the most advanced security tools, data sources, online databases and our own proprietary algorithms, which is exactly what makes HiveBot the most advanced product of its kind!</p><h3>Security Features</h3><p>As mentioned, every file goes through a long series of tests before returning a result.</p><p>For maximum security, the HiveBot is able to scan those files and links even <strong>before you download them</strong>, to keep you as safe as possible.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dCi_Qr9kZQLBO506mnzftQ.png" /><figcaption>HiveBot — @ zkHiveBot can scan every file without asking you to download it!</figcaption></figure><p>Following is a list of steps every PMC goes through to provide the scan results:</p><h4>1. Anti-Virus Checks</h4><p>Every file is scanned by multiple reputable AV vendors that have partnered with zkHive, including Kaspersky, ESET, Bitdefender &amp; more.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/914/1*CKjpoWKCF_5N0IAfK1RIlQ.png" /><figcaption>Anti Viruses Results</figcaption></figure><h4>2. Anti-Malware Checks</h4><p>Every file is scanned by multiple reputable AV vendors that have partnered with zkHive, including VBA32, K7 Security &amp; more.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/794/1*tVWaKbIWapZuzO2vGB6Qag.png" /><figcaption>Anti Malware Results</figcaption></figure><h4>3. Online Phishing Databases</h4><p>Our backend regulary monitors and downloads lists of well-known harmful files, links and hashes to make sure our bot is always up-to-date and provide the most accurate data available</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/664/1*yFnvgNSSuWRq27h2IO-BBA.png" /><figcaption>Phishing Databases</figcaption></figure><h4>4. Cloud Risk Providers</h4><p>HiveBot leverages cloud risk providers, such as Google Cloud Provider (GCP), which are reputable cloud companies that aggregate scans and reports from users all around the world on a daily basis.</p><h4>5. HiveMind</h4><p>This security analysis uses data scraped from the web, our security partners, user reports and all other sources we have, and <strong>feeds it into our proprietary AI and ML based algorithms.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/730/1*ZJaU4t33qHxwvcO_0imYTw.png" /></figure><h4>6. User Reports</h4><p>lastly, every PMC sent to the HiveBot can be flagged by every user as legit or suspicious, providing the <strong>largest community-built knowledge base for potential harmful content on Telegram.<br></strong>This is a <strong>key feature that provides the most prolific security assessment</strong> possible, because all other sources might take hours if not days to properly update, whereas the HiveBot can provide <strong>instantaneous user reports, </strong>the second they are happening</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/688/1*dqkIRuOu3nnE--RD-LeQqw.png" /><figcaption>User Reports</figcaption></figure><p>The full report should look like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/738/1*Tkor262n5iZ7hB6k8jeWRg.png" /><figcaption>Full report</figcaption></figure><h3>Start using it today!</h3><p>Just go to Telegram, open @ zkhiveBot and forward your files. The Hive will do the rest!</p><p>@ zkhiveBot — ON TELEGRAM</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=22ef301f159b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[zkHive: The Revenue Model]]></title>
            <link>https://medium.com/@zkhive/zkhive-the-revenue-model-87296059664b?source=rss-21b4d1e86838------2</link>
            <guid isPermaLink="false">https://medium.com/p/87296059664b</guid>
            <dc:creator><![CDATA[zkHive]]></dc:creator>
            <pubDate>Sat, 09 Mar 2024 11:51:56 GMT</pubDate>
            <atom:updated>2024-03-09T11:51:56.771Z</atom:updated>
            <content:encoded><![CDATA[<h4>A Comprehensive Analysis of the Hive’s revenue Generation Strategies</h4><p>With our upcoming beta launch next week it is now the perfect time to discuss the platform’s business model, revenue streams and how $ZKHIVE is used to take full advantage of the 1B$ opportunity zkHive is going to solve.</p><h3><strong>The Anti-Virus Market</strong></h3><p>The antivirus software market has been showing steady growth over recent years. In 2023, the market was valued at approximately $4.09 billion and is projected to grow nearly 50% by 2030.<br> This growth is attributed to factors like the rise in cybersecurity threats and increased internet connectivity.</p><p>This growth trajectory is driven by emerging threats such as ransomware, advanced persistent threats, IoT security concerns, and the integration of antivirus solutions with endpoint security.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lQGYR5wVmGz6sOQ343WWNg.png" /><figcaption>Source: <a href="http://market.us/">market.us</a></figcaption></figure><h3><strong>The Online Firewall Market</strong></h3><p>This sector shows an even faster size and growth, with nearly 6B$ in size in 2022 and an expected growth of more than 100% to $14 Billion by 2028.</p><p>Companies like Fastlify, CloudFlare and AWS are some of the fastest growing companies in the world, providing essential services for millions of customers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YCa10U3nnxf90zDul8udug.png" /><figcaption>WAF Marketcap Overview. Source: marketsandmarkets.com</figcaption></figure><h3><strong>The zkHive Opportunity — First Mover Advantage</strong></h3><p>Combining these market stats with our initial blogpost that explained how the specific market of social exploits and scams in crypto generates in excess of 1B$ of losses per year, is zkHive’s true business opportunity.</p><p>This first-mover advantage in the Telegram and twitter firewall space positions Hive not just as a trendsetter, but as a potential market leader in an area with escalating demand. As Telegram continues to grow as a preferred platform for communities, the need for robust security measures like Hive becomes increasingly critical.</p><p>This represents a significant market opportunity, tapping into both the extensive user base of Telegram and the broader push towards enhancing digital security.</p><h3><strong>$ZKHIVE revenue streams</strong></h3><p>Our native token, $ZKHIVE, is the heart of the entire ecosystem, designed to maximize the advantages for all the token holders. <br>With a comprehensive suite of benefits, $ZKHIVE token holders are poised to capitalize on zkHive’s immense revenue potential, estimated at $1 billion opportunity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5bMSuwNk6jXp8furbkhPTg.png" /><figcaption>zkHive revenue streams</figcaption></figure><h4>One-Time Fees — Installation fee</h4><p>First off, every user of our Decentralized Firewall will need to pay a one-time fee for the installation. The installation is seamless, just add the bot, pay the fee in an automated way and you are good to go!</p><p>A large part of this fee will go directly to the liquidity and token holders, and the rest will be used to cover development costs.</p><p>This business model has proved to be very useful, with well known projects such as Group Help, PAAL, AIX and others selling tens of millions of licenses.</p><h4>Premium Features</h4><p>The next utility of $ZKHIVE is gated feature access — some of our most advanced features such as our real-time threat intelligence platform will require the holders to pay with $ZKHIVE which will be burned out of circulation.</p><h4>zkHive For Individuals</h4><p>zkHive also provides great value to individuals, and not just communities by providing single-tap file and link scanning, decentralised reports and feedbacks, creating the largest user-based security network for individuals.</p><p>Users that would want to take full advantage of our security network would need to hold the $ZKHIVE token.</p><h4>API Access</h4><p>As zkHive’s security network grows, it will become an invaluable resource that other security products will be able to use.<br>These resources will be provided to 3rd party companies for a monthly fee, generating another stable revenue stream for the project.</p><h4>ADS Revenue</h4><p>With zkHive bot poised to operate in thousands of groups, substantial advertising revenue is anticipated. This broad reach ensures that our platform remains highly attractive to advertisers, further increasing our revenue streams and solidifying our position as a leader in the web3 security industry.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=87296059664b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Implementing AI Fraud Detection in the Crypto Space]]></title>
            <link>https://medium.com/@zkhive/implementing-ai-fraud-detection-in-the-crypto-space-9c9629cd1a35?source=rss-21b4d1e86838------2</link>
            <guid isPermaLink="false">https://medium.com/p/9c9629cd1a35</guid>
            <category><![CDATA[fraud-detection]]></category>
            <category><![CDATA[zero-knowledge]]></category>
            <category><![CDATA[ethereum]]></category>
            <category><![CDATA[zero-knowledge-proofs]]></category>
            <category><![CDATA[defi]]></category>
            <dc:creator><![CDATA[zkHive]]></dc:creator>
            <pubDate>Fri, 08 Mar 2024 13:03:43 GMT</pubDate>
            <atom:updated>2024-03-08T13:03:43.471Z</atom:updated>
            <content:encoded><![CDATA[<p>In the last post we introduced the major pain point zkHive is here to solve — advancing Telegram, Discord and Twitter security space, in order to fight the ever growing number of hacks and exploits resulting in nearly <strong>1 BILLION $ per year.</strong></p><p>In this post, we will dive into some of the research papers that are used by the zkHive team to build our AI-assisted decentralised peer-to-peer security network, that can stop these threats <strong>before </strong>they impact you as a group or as an individual</p><h3>Background</h3><p>First off, we would like to thank Massimo La Morgia et al (2023), BINGQIAO LUO et al (2023), Mary Isangediok and Kelum Gajamannage (2022).<br>Their research paper and others have been an invaluable source of knowledge for us in the challenge of integrating AI and ML into a cohesive and effective fraud detection solution.</p><h3>Introduction</h3><p>The integration of Artificial Intelligence (AI) in fraud detection is revolutionizing how financial institutions and businesses tackle security challenges. AI’s ability to analyze vast amounts of data quickly and accurately makes it an invaluable tool in identifying fraudulent activities that might otherwise go unnoticed.</p><p>In the first stage, AI systems gather and analyze data, using machine learning algorithms to detect patterns and anomalies indicative of fraudulent behavior. This process includes scrutinizing transaction histories, user behavior, and other relevant data points. AI’s predictive capabilities enable it to identify potential fraud before it occurs, a significant advantage over traditional methods that often detect fraud only after the fact.</p><p>The second stage involves the continuous learning and adaptation of AI models. As they are exposed to new data and fraud tactics, these models evolve, becoming more sophisticated and accurate in their predictions. This adaptability is crucial in the ever-changing landscape of cyber threats and financial scams.</p><p>With the help of the papers mentioned above, we used an extensive examination of Telegram, analyzing data from over 120,000 public channels and more than 200 million messages, with the focus on identifying and studying fake channels and fraudulent activity within the platform. Using their proposed machine learning model, it is possible to detect these fake channels, achieving an impressive <strong>85% accuracy rate</strong>.</p><h3>Stats on DeFi Attacks and Fraudulent Activity</h3><p>Figure 1 illustrates that from the first major cryptocurrency exchange hack in 2011 (Mt. Gox) through February 2023, there has been a notable escalation in DeFi fraud. This period has seen a total of 231 DeFi hacks, 135 security breaches, and 95 fraudulent activities, cumulatively leading to an estimated loss of about $16.7 billion in cryptocurrencies over a 12-year span.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/890/1*dQFY86ihkYtF2nTlwDLwfQ.png" /><figcaption>Fig. 1: Number and amounts of DeFi attacks. Data from [1]</figcaption></figure><p>Researchers have explored different methods to detect phishing websites and giveaway scams in DeFi. Phillips and Wilder [3] employed DBSCAN clustering [4] for identifying advance-fee and phishing scams, analyzing data from social media and blockchain. Li et al. [5] introduced CryptoScamTracker, a system that uses Certificate Transparency Logs for scam detection. They also used the AllenNLP model [6] in CryptoScamHunter to identify Twitter giveaway scams. Another study [7] utilized machine learning techniques like Decision Trees and SVMs to detect phishing URLs in fraudulent NFT projects on Twitter. Despite these advancements, there’s still a need for machine learning techniques that incorporate on-chain data in scam detection.</p><p>Figure 2 is a schematic of TTANG, which is a leading work for phishing detection using graphrelated techniques. It comprises three modules: transaction graph generation, network embedding learning, and phishing address detection.<br>The inner workings of the model are beyong the scope of this article, but we encourage all readers to dive deeper to the links below (esp. [2] for more info on TTANG itself).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ul6L3fta6Ke65ACQzUHbEg.png" /><figcaption>Fig. 2: Process of Detecting Phishing Scams. Adapted from [2]</figcaption></figure><p>Lastly we will explore the issue of fake socials on Telegram. Do note that while the data focuses on Telegram, the observations and metrics are applicable to other social networks with similar characteristics such as Discord.</p><p>Figure 3 presents different simple characteristics of different group types — fake groups, official groups and the provided TGDataset aggregate, which is one of the largest general purpose snapshot of the TG ecosystem built by the researchers at [8].</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*i_n1RCupxvaaTGYtEc1Wrw.png" /><figcaption>Fig. 3: Telegram stats by group aggregation types — fake, official and standard. Source: [8]</figcaption></figure><p>Fig. 3 is especially powerful because it shows stark differences between legitimate and fraudulent activities on Telegram. For example, the high forwarded messages ratio produced in fake channels is a strong classifier compared to the low ratio of legitimate groups. Another example is the high number of media messages in fake groups compared to official ones.</p><h3>Optimized Machine Learning vs Fraud</h3><p>In this section we will shed some light on several machine learning tools and methods that are used to detect fraudulent activity.<br>Once again, the goal of this article is to provide high level overview. Interested readers are encouraged to research the subject using the provided bibliography.</p><p>The 2 main classifiers we are interested at are logistic regression and extreme gradient boosting (or XGBoost).</p><p><strong>Logistic regression (LR)</strong>, known for its straightforwardness [9], is a prevalent algorithm in binary classification tasks. In the context of fraud detection, LR is used to model the probability of a transaction being fraudulent. It does this by designating fraud cases as 1s and legitimate transactions as 0s, using a set threshold to determine the likelihood based on various features present in the dataset.</p><p>the logistic function is given as</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/794/1*lwmivFRq0s_z69xSWXDRzA.png" /></figure><p>the LR model given the probability fw,b(x) is</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/832/1*xZ7XyaX8o0w5585SAk3Owg.png" /></figure><p>where x denote the input feature vector of length m, and the response z is given as a straight line z = w · x + b, where w is the weights and b is the bias term estimated during training.</p><p>LR is reliable and requires less computation time for the training phase, at the cost of inaccuracy in some cases.</p><p><strong>Extreme Gradient Boosting</strong>, also known as XGBoosting (XGB) [10], enhances predictive accuracy by constructing a series of trees. It utilizes a method called sampling with replacement to generate new training datasets. This approach is key to its ability to boost the performance of predictive models.</p><p>The input yi can be computed as follows:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/502/1*QARTD7AwWfEK059EOIY_Sg.png" /></figure><p>Where F is the space of regression trees and each fk corresponds to an independent tree structure.<br>This can be used to greedily calculate the objective L</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/696/1*v5rJqabXFjX-U05iobrr5Q.png" /></figure><p>XGB can produce better results and much more customizability using all of its hyperparameters, at the cost of high computation time is the biggest disadvantage of this algorithm.</p><h3>Summary</h3><p>In conclusion, AI’s role in fraud detection is not only about identifying current fraudulent activities but also about predicting and adapting to future threats. Its integration offers a more proactive approach, shifting the focus from reaction to prevention. The use of AI in this field is expected to grow, becoming an essential component of financial security strategies.</p><p>We shared some stats and datasets our product is based upon, and 2 of the machine learning methods that are used for fraud detection — LR which is fast but may be inaccurate and XGB which is more powerful but more costly.</p><p>For more comprehensive insights into specific use cases and technologies, it’s recommended to explore specialized literature and research papers in the field of AI and cybersecurity.</p><h3>Socials</h3><p>TG: zkHive_token<br>Twitter: zkHive<br>website: zkhive . io</p><h3>Bibliography:</h3><p>[1] Crystal. 2023. Crypto &amp; DeFi Security Breaches, Fraud &amp; Scams Report. <a href="https://crystalblockchain.com/security-breaches-and-fraud-involvingcrypto/">https://crystalblockchain.com/security-breaches-and-fraud-involvingcrypto/</a> Accessed on Aug 1, 2023.</p><p>[2] Sijia Li, Gaopeng Gou, Chang Liu, Chengshang Hou, Zhenzhen Li, and Gang Xiong. 2022. TTAGN: Temporal transaction aggregation graph network for ethereum phishing scams detection. In Proceedings of the ACM Web Conference 2022. 661–669.</p><p>[3] Ross Phillips and Heidi Wilder. 2020. Tracing cryptocurrency scams: Clustering replicated advance-fee and phishing websites. In 2020 IEEE International Conference on Blockchain and Cryptocurrency (ICBC). IEEE, 1–8.</p><p>[4] Dingsheng Deng. 2020. DBSCAN clustering algorithm based on density. In 2020 7th international forum on electrical engineering and automation (IFEEA). IEEE, 949–953.</p><p>[5] Xigao Li, Anurag Yepuri, and Nick Nikiforakis. 2023. Double and Nothing: Understanding and Detecting Cryptocurrency Giveaway Scams. In Network and Distributed Systems Security (NDSS) Symposium.</p><p>[6] Matt Gardner, Joel Grus, Mark Neumann, Oyvind Tafjord, Pradeep Dasigi, Nelson Liu, Matthew Peters, Michael Schmitz, and Luke Zettlemoyer. 2018. Allennlp: A deep semantic natural language processing platform. arXiv preprint arXiv:1803.07640 (2018).</p><p>[7] Sayak Saha Roy, Dipanjan Das, Priyanka Bose, Christopher Kruegel, Giovanni Vigna, and Shirin Nilizadeh. 2023. Demystifying NFT Promotion and Phishing Scams. arXiv e-prints (2023), arXiv–2301.</p><p>[8] It’s a Trap! Detection and Analysis of Fake Channels on Telegram Massimo La Morgia1 , Alessandro Mei1 , Alberto Maria Mongardini1 , and Jie Wu2</p><p>[9] Sandro Sperandei. Understanding logistic regression analysis. Biochemia medica, 24(1):12–18, 2014.</p><p>[10] Tianqi Chen and Carlos Guestrin. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, pages 785–794, 2016.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9c9629cd1a35" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[We Are zkHive: The First AI-assisted Decentralized Firewall For Crypto]]></title>
            <link>https://medium.com/@zkhive/we-are-zkhive-the-first-ai-assisted-decentralized-firewall-for-crypto-9c61a98d7529?source=rss-21b4d1e86838------2</link>
            <guid isPermaLink="false">https://medium.com/p/9c61a98d7529</guid>
            <category><![CDATA[eth]]></category>
            <category><![CDATA[zero-knowledge-proofs]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[fraud-detection]]></category>
            <dc:creator><![CDATA[zkHive]]></dc:creator>
            <pubDate>Wed, 06 Mar 2024 12:42:56 GMT</pubDate>
            <atom:updated>2024-03-06T12:42:56.882Z</atom:updated>
            <content:encoded><![CDATA[<h4>Redefining Security With Connectivity AI</h4><blockquote>“Just like the working bees in the hive work together to keep each other safe, zkHive combines the knowledge of thousands of groups and hundreds of thousands of users to keep the entire network safe” — zkHive CEO</blockquote><p>zkHive is an AI-assisted formal verification platform, that leverages a zero-knowledge decentralized peer-to-peer security network to keep crypto communities and individuals safe.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dkkl6eUiD7yYU8gXOLAXTA.png" /><figcaption>zkHive is addressing the biggest fraud attacks in crypto. Image credit: Quill audits.</figcaption></figure><p>In this article we will explain the motivation behind zkHive, and the $1B issue it solves.</p><h3>The Problem</h3><p>Telegram, Twitter and Discord have become the de-facto city square for crypto enthusiasts, containing more than 70% of crypto-related launches and discussions.</p><p>However, their security state is at an all-time low, with <strong>thousands of hacks and exploits per month</strong>, resulting in losses of more than <strong>500,000,000$</strong> <strong>per year</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6w-2RTPpfXNSLGV4WleMUw.png" /><figcaption>During the last bull market, Crypto related frauds surpassed frauds in every other industry combined</figcaption></figure><p>Existing community protection solutions are based on <strong>early 2010’s tech</strong>: human mods, captcha on join and simple message counting (anti flood).</p><p><strong>This is in stark contrast</strong> to Web2 security that has active monitoring, 24/7 analysis of behaviour and traffic</p><p>For individual users, the situation is even more dire. Users are getting scammed every day because of malicious GIFs, files, scams, links, PDFs etc.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nwmtSqauIWUBPx8fsslLIA.png" /><figcaption>Social engineering is one of the most profitable attack vectors, especially on Telegram and Twitter</figcaption></figure><h3>The Solution: A connected, Decentralized Network</h3><p>Powered by connectivity AI, zkHive is a state-of-the-art security product that constantly filters out potentially malicious activity in order to keep crypto communities free of fraudulent content, bots and malicious actors before they get a chance to act.</p><p>zkHive is the first platform that protects users from 0-day and 1-day attacks thanks to our proprietary zero-knowledge based technology.<br>Dozens of platforms secure code, but only zkHive secures people.</p><p>zkHive synergizes well with all existing security solutions, harnessing the power of the Peer-To-Peer AI in your group.</p><h3>Features</h3><p>zkHive is packed with groundbreaking features that will be announced soon, but here is a small glimplse of what it offers.</p><ul><li><strong>Real-time Threat Detection</strong> — <br>Leverage the connectivity AI capabilities to monitor user activity across all groups, preemptively detecting and preventing bot attacks, organized FUD, and account takeovers before they compromise your community’s security, ensuring constant vigilance and protection.</li><li><strong>AI-Optimized Comprehensive File and Message Scanning</strong> — <br>Seamlessly scan all files, links, GIFs, documents, and voice messages in Telegram, providing 365-day protection against malicious content and potential scams.</li><li><strong>Leverage The Decentralized Hive Mind — <br></strong>Harness the power of the first ever decentralized security network, built and evolves using AI peer-to-peer technology!Whenever a threat is detect in any of zkHive’s groups or reported by individuals, it is immediately blocked in all of them.</li><li><strong>AI-Optimized Specialized Security Modes — <br></strong>Real-Time Under Attack mode for Telegram Based on well known Big-Tech “Under Attack” modes, zkHive allows you to immediately activate this special mode when being attacked, which would immediately deeply analyse all user activities and leverage our immense Database to squash any attack in no time!</li></ul><h3>Harnessing Machine Learning For Fraud Detection</h3><p>These features are possible thanks to the work of great researchers in the field of fraud analysis and AI-assisted malicious behaviour detection.</p><p>One example of those researches’ accomplishments is the use of Logistic Regression for binary classification of tasks.</p><p>In the context of fraud detection, LR is used to model the probability of a transaction being fraudulent.<br>Here’s a logistic function that could be used for this purpose that was analysed by Mary Isangediok and Kelum Gajamannage in their September 2022 paper</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/794/1*-eJcnpqgL1h4ryENhQco-A.png" /></figure><p>Which then gives the probability using the following formula:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/832/1*vm-MtoJbDRfb_jyEF6rNgA.png" /></figure><p>Seeing as this is the first blog post we don’t want to make it too technical, but the next post will cover these topics at greater depths.</p><p>In the next series of posts, we will cover the major features zkHive has implemented, the academic research it is based upon and how it is going to reduce scams and exploits by up to 90%, resulting in more than 5<strong><em>00M$</em></strong> of saved funds per year.</p><h4>Follow Us</h4><p>TG: zkHive_token<br>Twitter: zkHive<br>GitHub: zkHive<br>Website: zkHive . io</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9c61a98d7529" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>