<ul class="summary-list"><li><keyword id="11892960" type="UNKNOWN" weightage="60" keywordseo="facebook" source="Orion">Facebook</keyword> <keyword id="12110043" type="UNKNOWN" weightage="60" keywordseo="whistleblower" source="Orion">whistleblower</keyword> <keyword id="12236552" type="UNKNOWN" weightage="60" keywordseo="frances-haugen" source="Orion">Frances Haugen</keyword> leaked a trove of documents that Insider has since reviewed.</li><li>The records provide an inside look at the tech giant's lexicon, from "harass dash" to "eat your veggies."</li></ul><p><em>These Facebook terms and definitions are sourced from disclosures made to the Securities and Exchange Commission and provided to Congress in redacted form by Frances Haugen's legal counsel. A consortium of <keyword id="11893589" type="UNKNOWN" weightage="60" keywordseo="news" source="Orion">news</keyword> organizations, including Insider, obtained the redacted versions received by Congress. </em></p><p>The majority of this list is sourced from the internal glossary used by Facebook's integrity team. Other terms were found within various company presentations and employee posts included in the Facebook Papers. Facebook did not respond to Insider's request for comment. </p><h3>1. Eat your veggies</h3><p>Code word for Facebook's communications protocol surrounding "major/sensitive news feed launches."</p><h3>2. Cannibalism</h3><p>A general business concept that describes how the growth of one company product may have negative implications on the growth of other company products.</p><p>Facebook employees are told to consider "incrementality:" to what degree is the growth of one product, such as Instagram, considered "truly incremental growth" when compared to the growth of other products, such as the main Facebook app. </p><h3>3. Dark social</h3><p>Term for private encrypted social messaging such as WhatsApp. Misinformation shared on dark social is "hard to detect, track, or block."</p><h3>4. Harass dash</h3><p>Short for an internal tool called "the harassment dashboard." The harass dash uses classifiers such as "bullying and harassment," "hate speech," and "violence and incitement" to flag comments that may violate Facebook's community standards.</p><h3>5. Waterfall </h3><p>The "primary production pipeline for misinformation."</p><h3>6. Surge</h3><p>2020 was "the year of the surge," according to an internal Facebook presentation called "Working Sustainably in 2021 (and beyond)." One slide defined surges as "periods of intense refocusing and doubling down on a set of products, policies, and processes intended to meet the moments we found ourselves in." These "moments" have included the pandemic, racial justice, economic distress, and the presidential election. </p><h3>7. Lockdown </h3><p>When a Facebook team is "put into lockdown" to solve a problem, it is a "time-bounded, tightly organized processes with clear ownership and one specific goal." Due to the high-priority status of a lockdown, the affected team "may request additional resources from neighboring teams." </p><p>The author of the presentation defining terms such as surge, lockdown, and war room, noted: "... many of our work terms have military connotations - while there is much to admire in the rigorous discipline and courage of these institutions, words matter. Using the language of battle likely subconsciously impacts the lens with which we view our mission and our output."</p><h3>8. Astroturf</h3><p>Facebook groups that are initially populated by fake accounts, but become self-sustaining as real accounts become participants. </p><h3>9. Blackhole</h3><p>A blacklisting service for URLs, domains, and IP addresses that may distribute spam or bad content. Facebook uses Blackhole<strong> </strong>to attach a label such as "OK," "BAD," or "IFFY" to these objects, which has a corresponding effect on how it can be used and viewed on the site.</p><h3>10. FAIR</h3><p>Acronym for Facebook AI Research. </p><h3>11. Remove, Reduce, Inform</h3><p>Facebook's framework for fighting misinformation on the integrity team: "we remove the content that violates our policies, reduce the spread of problematic content that does not violate our policies, and provide additional context to help people make informed choices about what they read or share."</p><h3>12. Depolarizing group</h3><p>A Facebook Group that enables positive interactions between people who are far apart on the ideological spectrum. </p><h3>13. CORGI</h3><p>Extremely complex math that looks for patterns in user behavior and identifies cliques on the platform -"very useful" for identifying large-scale attempts at manipulation. </p><h3>14. Your money, your life  </h3><p>The acronym "YMYL" may refer to posts, pages, websites, or apps which touch on issues affecting a person's livelihood or welfare. "Spam or abuse on such topics can be especially harmful," according to Facebook's integrity team glossary.</p><h3>15. Faux-tire</h3><p>Refers to fake satire: "material meant to misinform/push propaganda while trying to self-portray as satire to avoid classification as newsworthy," thus avoiding fact checks. </p><h3>16. Magnet users</h3><p>Facebook users who are "hyper-engaged with bad content."</p><h3>17. Mystery image</h3><p>Clickbait images that may withhold information "involving superimposed synthetic graphics that put unnatural attention to parts of an image. These are typically low-quality, spammy user experiences."</p><h3>18. Sacramento</h3><p>Facebook's code name for "an A/C priv data project." This likely refers to "attorney-client privilege," a legal rule protecting confidential communication between an attorney and his or her client from discovery processes. </p><h3>19. Speaking </h3><p>The code name for the "approved talking points" of Facebook's integrity team: "In Integrity, be careful to always stay inside our department's approved talking points ... due to the sensitivity of our topics always preflight any possible messaging through appropriate channels."</p><h3>20. Critical events</h3><p>The term used by Facebook's integrity team to reference events such as "elections, crises, and chronic social problems." Critical events may require "operations and tooling to provide an extra focus on proactive sweeps and manually trigger actions to mitigate these highest risk episodes." </p><h3>21. Divisive content</h3><p>Facebook's integrity team defines divisive content as "content that ideologically similar people have positive responses to and ideologically different people have negative responses to."</p><h3>22. False news</h3><p>Facebook now labels "fake news" as false news, calling the former term for misleading public information "politicized."</p><h3>23. Aligned friend and Cross-cutting friend</h3><p>An aligned friend is a person who is near you in the ideological map, also known as "in-group," while a cross-cutting friend is a person who is far from you on the ideological map, or in the "out-group."</p><h3>24. Blame tool</h3><p>An internal tool that provides data on "which sources and triggers recommended a bad post."</p><h3>25. Blue</h3><p>An adjective referring to the main Facebook app. For example, "blue time" would refer to the total time spent on the main app.  </p>

From 'harass dash' to 'dark social,' 25 Facebook code words revealed in leaked whistleblower documents - and what they all mean

<keyword id="790242" type="COMPANY" weightage="20" keywordseo="WhatsApp">WhatsApp</keyword> banned two million Indian accounts while it received 345 grievance reports between May 15 and June 15, the company said in its maiden monthly compliance report as mandated by the IT rules.<br /><br />The new IT rules require large digital platforms -- with over five million users -- to publish compliance reports every month, mentioning the details of complaints received and action taken.<br /><br />"Our top focus is preventing accounts from sending harmful or unwanted messages at scale. We maintain advanced capabilities to identify these accounts sending a high or abnormal rate of messages and banned two million accounts in India alone from May 15-June 15 attempting this kind of abuse," WhatsApp said on Thursday.<br /><br />WhatsApp clarified that more than 95 per cent of such bans are due to the unauthorised use of automated or bulk messaging (spam).<br /><br />"We expect to publish subsequent editions of the report 30-45 days after the reporting period to allow sufficient time for data collection and validation," WhatsApp said in the report.<br /><br />The Facebook-owned company explained that the number of accounts banned have risen significantly since 2019 as the sophistication of systems has increased, and "so we are catching more accounts even as we believe there are more attempts to send bulk or automated messages".<br /><br />The vast majority of these accounts are banned proactively, without relying on any user reports, it added.<br /><br />About eight million accounts are banned/disabled globally on an average per month.<br /><br />WhatsApp said besides the behavioural signals from accounts, it relies on available "unencrypted information" including user reports, profile photos, group photos and descriptions as well as advanced AI tools and resources to detect and prevent abuse on its platform.<br /><br />In the report, WhatsApp said it had received 345 reports in total, cutting across categories such as ban appeal, account support, product support, safety issues and others.<br /><br />Against this, 63 accounts were "actioned" by WhatsApp during May 15-June 15, 2021.<br /><br />WhatsApp said user reports received by the platform via the grievance <keyword id="11650763" type="UNKNOWN" weightage="20" keywordseo="channel/s">channel/s</keyword> are evaluated and responded to.<br /><br />Majority of users who reach out to WhatsApp are either aiming to have their account restored following an action to ban them or reaching out for product or account support, it added.<br /><br />'Accounts Actioned' denotes reports where WhatsApp took remedial action based on the report. Taking action denotes either banning an account or a previously banned account being restored as a result of the complaint.<br /><br />The IT rules -- which came into effect on May 26 -- mandate that significant digital platforms include the number of specific communication links or parts of information they proactively remove by using automated tools.<br /><br />Other platforms like <keyword id="6355629" type="DEFINITION" weightage="20" keywordseo="google">Google</keyword>, Koo and <keyword id="6354616" type="UNKNOWN" weightage="20" keywordseo="Twitter">Twitter</keyword> have already submitted their compliance reports. Instagram and <keyword id="6354576" type="UNKNOWN" weightage="20" keywordseo="facebook">Facebook</keyword> have also submitted their reports.<br /><br />The IT rules have been designed to prevent abuse and misuse of digital platforms, and offer users a robust forum for grievance redressal. Under these rules, social media companies will have to take down flagged content within 36 hours, and remove within 24 hours content that is flagged for nudity and pornography.<br /><br />The rules also mandate appointment of three key personnel -- grievance officer, chief compliance officer and nodal officer. These officials need to be residents in India. Non-compliance with the IT rules would result in these platforms losing their intermediary status that provides them immunity from liabilities over any third-party data hosted by them.<br /><br /><strong>SEE ALSO:<br /><a target="_blank" class="c-link c-message_attachment__title_link" href="https://www.businessinsider.in/stock-market/news/tatva-chintan-pharma-chem-ipo-heres-how-to-apply-via-banks-and-apps-like-zerodha-upstox/articleshow/84462586.cms">Tatva Chintan Pharma Chem IPO: Here’s how to apply via banks and apps like Zerodha, Upstox</a><br /></strong><a target="_blank" class="c-link c-message_attachment__title_link" href="https://www.businessinsider.in/careers/news/wipro-hires-12000-employees-in-june-quarter-highest-in-decade/articleshow/84445858.cms"><strong>Wipro's hiring at highest in a decade to hedge against employees leaving</strong></a><br />

WhatsApp banned two million Indian accounts during May 15-June 15 period

<ul><li><strong>The "attack campaign" of this 'Trojan' category virus is active globally, said the Computer Emergency Response Team of India (CERT-In), the national technology arm to combat cyberattacks and guard Indian cyber space. </strong></li><li><strong>"It is reported that a new Android <keyword id="54201" type="PERSON" weightage="20" keywordseo="Malware">malware</keyword> strain dubbed 'BlackRock' equipped with datastealing capabilities is attacking a wide range of Android applications. </strong></li><li><strong>The "noteworthy feature" of this malware is that its target list contains 337 applications including banking and financial applications, and also non-financial and well-known commonly used brand name apps on an <keyword id="747689" type="PERSON" weightage="20" keywordseo="Android-device">Android device</keyword> that focus on social, communication, networking and dating platforms, it said. </strong></li></ul>The country's cyber security agency has issued an alert against an Android malware, dubbed 'BlackRock', that has the potential to "steal" banking and other confidential data of a user. It can extract credentials and credit card information from over 300 apps such as email, e-commerce apps, social media apps, besides banking and financial apps, the CERT-In said in an advisory.<br /><br />The "attack campaign" of this 'Trojan' category virus is active globally, said the Computer Emergency Response Team of India (CERT-In), the national technology arm to combat cyberattacks and guard Indian cyber space. <br /><br />"It is reported that a new Android malware strain dubbed 'BlackRock' equipped with datastealing capabilities is attacking a wide range of Android applications. <br /><br />"The malware is developed using the source code of Xerxes banking malware which itself is a variant of LokiBot Android Trojan," the advisory said. <br /><br />The "noteworthy feature" of this malware is that its target list contains 337 applications including banking and financial applications, and also non-financial and well-known commonly used brand name apps on an Android device that focus on social, communication, networking and dating platforms, it said. <br /><br />"It can steal credentials and credit card information from over 300 plus apps like email clients, e-commerce apps, virtual currency, messaging or social media apps, entertainment apps, banking and financial apps etc," the advisory said. <br /><br />The advisory described the infection activity of the virus. <br /><br />"When the malware is launched on the victim's device, it hides its icon from <keyword id="3260481" type="COMPANY" weightage="20" keywordseo="app-drawer">app drawer</keyword> and then masquerades itself as a fake <keyword id="6355629" type="COMPANY" weightage="20" keywordseo="google">Google</keyword> update to request accessibility service privileges."<br /><br />"Once this privilege is granted, it becomes free to grant itself additional permissions allowing it to function further without interacting with user," it said. <br /><br />Threat operators can issue a number of commands for various operations such as logging keystrokes, spamming the victims' contact lists with text messages, setting the malware as the default SMS manager, pushing system notifications to the C2 (command and control) server, locking the victim in the device home screen and steal and hide notifications, send spam and steal SMS messages and many more such activities, the advisory said. <br /><br />The virus is deadly as it has the capability to "deflect" majority of anti-virus applications. <br /><br />"Another feature of this Android Trojan is making use of "Android work profiles" to control the compromised device without requiring complete admin rights and instead creating and attributing its own managed profile to gain admin privileges," it said. <br /><br />The federal cyber security agency suggested some counter-measures: do not download and install applications from untrusted sources and use reputed application market only; always review the app details, number of downloads, user reviews and check 'additional information' section before downloading an app from play store, use device encryption or encrypt external SD card; avoid using unsecured, unknown Wi-Fi networks among others. <br /><br />Also, when it comes to downloading banking apps one should use the official and verified version and users should make sure they have a strong AI-powered mobile antivirus installed to detect and block this kind of tricky malware, the advisory said.  <br /><br />SEE ALSO:<br /><a href="https://www.businessinsider.in/business/news/reliance-retail-contribution-to-ril-as-mukesh-ambani-preps-for-fundraising-will-be-key/articleshow/77240105.cms">Reliance Retail’s contribution to RIL as Mukesh Ambani preps for fundraising will be key⁠— aside from the oil shock</a><br /><a href="https://www.businessinsider.in/education/news/india-revamps-its-education-policy-after-34-years-scraps-mphil-proposes-vocational-courses-from-class-6-ug-colleges-to-become-autonomous/articleshow/77241851.cms">India revamps its education policy after 34 years — scraps MPhil, proposes vocational courses from class 6, UG colleges to become autonomous</a><br /><a href="https://www.businessinsider.in/education/news/india-revamps-its-education-policy-after-34-years-scraps-mphil-proposes-vocational-courses-from-class-6-ug-colleges-to-become-autonomous/articleshow/77241851.cms">India’s Rafale fighter jets — Here’s how it stacks up against Chinese J-20</a><br />

Android malware 'BlackRock' prowling in cyber space, may steal banking data: Advisory

<p><span style="font-weight: 400;"><div class="KonaFilter image-container display-table"><div><div class="image on-image"><img src="https://static-ssl.businessinsider.com/image/5d139093e3ecba024912413a-2400/kirsten wegner.jpg" border="0" alt="Kirsten Wegner" data-mce-source="Courtesy of Kirsten Wegner" data-mce-caption="Kirsten Wegner of MMI." /><p class="source"><span>Courtesy of Kirsten Wegner</span></p><p class="caption">Kirsten Wegner, CEO of the Modern Markets Initiative.</p></div><p></p></div></div></span></p><ul class="summary-list"><li><strong>Kirsten Wegner is CEO of the Modern Markets Initiative, an advocacy group advancing secure savings and investment through innovation.</strong></li><li><strong>In this opinion piece, she writes that for millions of Americans, and even many in Congress, the real uses of AI in the financial markets remains much of a mystery. </strong></li><li><strong>As Congress examines this important issue, it is useful to step back and look more broadly at trends toward automated functions in the financial markets and what AI means or should mean to market participants.</strong></li><li><strong>She outlines four points to consider, including optimism that AI will enable financial institutions to broaden access to banking and credit to traditionally underserved communities.</strong></li><li><strong><a href="https://www.businessinsider.com/?hprecirc-bullet" target="_blank" rel="noopener">Visit Business Insider's homepage for more stories.</a></strong></li></ul><p><span style="font-weight: 400;">Over the past few decades, life as many of us know it has changed dramatically due to the evolution of AI - artificial intelligence. Whether it's using AI to help humans decide which emails should get routed to a spam filter, or recommending music or movies based on past preferences, to voice dictation on your smart phone, rarely a day goes by where data science isn't analyzing huge amounts of information to help humans make more informed decisions.</span></p><p><span style="font-weight: 400;">But according to a recent HubSpot </span><a href="https://blog.hubspot.com/news-trends/artificial-intelligence-is-here"><span style="font-weight: 400;">survey</span></a><span style="font-weight: 400;">, almost two-thirds of consumers aren't aware that they are even using AI technologies. And for millions of Americans, and even many in Congress, the real uses of AI in the financial markets remains much of a mystery. On Wednesday of this week, the US House Financial Services Committee will hold a </span><a href="https://financialservices.house.gov/calendar/eventsingle.aspx?EventID=403824"><span style="font-weight: 400;">hearing</span></a><span style="font-weight: 400;"> on where we are and the next frontier on the uses of AI in the markets.</span></p><p><span style="font-weight: 400;">As Congress examines this important issue, it is useful to step back and look more broadly at trends toward automated functions in the financial markets and what AI means or should mean to market participants. Here are four key points to consider.</span></p><h2><strong>1. AI is a driver of economic efficiency and growth</strong></h2><p><span style="font-weight: 400;">Machine learning is a tool that allows humans to do things that analyze vast amounts of data and can lead to more informed decision making though its ability to analyze advanced data. The power of the human mind, assisted with the tool of machine learning, together can make decisions much more efficiently. </span></p><p><span style="font-weight: 400;">AI is a tool for providing data to humans on both qualitative and quantitative inputs, allowing humans to make better informed trade decisions. For example, machine learning can listen and analyze multiple earnings calls at once, interpret data, and provide more in-depth insight for analysts. </span></p><h2><strong>2. AI drives cost savings for investors by being more effective </strong></h2><p><span style="font-weight: 400;">AI as a tool for humans has the potential to make the markets more efficient, tighten the bid-ask spread, decrease processing times for transactions, more accurately monitor systems and behavior, and provide business analysis that is based on a greater depth of data - all at a cost savings to investors. It has the potential to dramatically lower operating costs for a firm, thus lowering fees to consumers.</span></p><p><em><span style="font-weight: 400;"><strong>Read more: </strong> <a href="https://markets.businessinsider.com/currencies/news/5-theories-bitcoin-price-tripled-to-12000-in-three-months-2019-6-1028310147">5 theories behind bitcoin's dizzying rally above $12,000 </a></span></em></p><h2><strong>3. AI and machine-learning are essential for improving market surveillance, cyber security, and fraud detection</strong></h2><p><span style="font-weight: 400;">AI's important potential applications in the cybersecurity sphere, both in designing and operating secure systems based on data, can supplement the time and attention of human experts. Automating market surveillance has the potential to yield greater security for financial transactions at decreased costs, as well as a more rapid response time.</span></p><p><span style="font-weight: 400;">Several high frequency trading firms are already working to use algorithms to detect potential bad actors in the market and are working on private-public partnerships with financial regulators such as FINRA to deploy artificial intelligence to surveil the markets.</span></p><h2><strong>4. Inclusion is an important factor to consider as AI evolves</strong></h2><p><span style="font-weight: 400;">From a workforce perspective, it seems inevitable that as society shifts toward greater automation that there will be a growth in jobs in the coding/technology industries. It is obligatory on the private sector, educators, and lawmakers to drive initiatives that attract diverse candidates to the workforce, so they too may benefit from the economic opportunities of AI and machine learning. </span></p><p><span style="font-weight: 400;">There are reasons for optimism that AI will enable financial institutions to broaden access to banking and credit to traditionally underserved communities, including the 20 percent of Americans who are currently unbanked or underbanked. There are new innovative business models evolving using responsible innovation including alternative data to, for example, replace credit scores with other data points to give the unbanked or underbanked communities access to credit. Further, data science can enhance "getting to know you" customer experiences to serve a broader array of communities.</span></p><p><span style="font-weight: 400;">Thanks to our innovation economy, the US enjoys the best markets for investors in terms of trade execution, cost of trading, savings delivered to investors, and secure and sound markets. The responsible use of AI has the opportunity to deliver even further efficient trading to investors, to give private sector and public sector new tools to detect fraud and abuse or cyberthreats, to promote inclusion initiatives, and to boost savings for Americans.</span></p><p><em><span style="font-weight: 400;">Kirsten Wegner is CEO of the Modern Markets Initiative, an advocacy group advancing secure savings and investment through innovation.</span></em></p><p> </p><p class="tagline">This is an opinion column. The thoughts expressed are those of the author(s).</p>    <div class="margin-top">        <h3>NOW WATCH: <a href="https://www.businessinsider.com/tesla-elon-musk-ceo-net-worth-makes-spends-billion-money-2019-6">How Tesla CEO Elon Musk makes and spends his $19.2 billion</a></h3>        <div>            <iframe class="bijwplayerclass" data-player="jwplayer" src="https://content.jwplatform.com/players/6oM6LxSR-CEZHWCBG.html" width="100%" height="100%" frameborder="0" scrolling="auto" style="margin:0px;" allowfullscreen=""></iframe><script>function receiveMessage(event){ console.log(event.data.fullscreen); Android.fullscreen(event.data.fullscreen); }window.addEventListener("message", receiveMessage, false);</script>        </div>    </div>

Responsible use of artificial intelligence can make our financial markets even stronger

<p><div class="KonaFilter image-container display-table"><div><div class="image on-image"><img src="https://static-ssl.businessinsider.com/image/5b07c7b21ae66244008b46c6-1116/screen shot 2016-03-19 at 31328 pm.png" border="0" alt="kim kardashian selfie" data-mce-source="Getty Images" /><p class="source"><span>Getty Images</span></p><p class="caption">Kim Kardashian has 111 million Instagram followers.</p></div><p></p></div></div></p><ul><li><strong>Instagram tracks exactly how you tap and scroll inside the app.</strong></li><li><strong>That's unnerving, but the company has an explanation.</strong></li><li><strong>Instagram has had a major bot problem, with accounts automatically liking and following users.</strong></li><li><strong>Tracking how humans behave on Instagram means the company is probably trying to use artificial intelligence to distinguish genuine accounts from bot behaviour.</strong></li></ul><hr /><p><br />Instagram is asking all its European users to agree to its updated terms and conditions, and anyone who goes through its current crop of notifications might notice one strange update: The company is tracking the way you tap and scroll through the app.</p><p>Under the new European General Data Protection Regulation (GDPR), Instagram and its parent Facebook must be clear about the kinds of information they collect about users, and why they're collecting it.</p><p>In its notification to users, Instagram explained: "We receive different kinds of information from your device, like how you tap and scroll, which can help distinguish humans from bots and detect fraud."</p><p><div class="KonaFilter image-container display-table"><div><div class="image on-image"><img src="https://static-ssl.businessinsider.com/image/5b07c51f1ae66230008b4d33-2400/instagrambots.png" border="0" alt="Instagram tap and scroll GDPR" data-mce-source="Shona Ghosh/Business Insider" /><p class="source"><span>Shona Ghosh/Business Insider</span></p></div><p></p></div></div></p><p>It appears that Instagram first notified users about the tracking in a policy update last month.</p><p>The idea that Instagram is keeping tabs on everything its users tap on is unnerving but, from the company's perspective, there is a good explanation.</p><p>Instagram has an acknowledged bot problem, and <a href="http://www.businessinsider.com/italian-security-researchers-find-8-percent-of-instagram-accounts-are-fake-2015-7">in 2015 one researcher estimated that 8% of all Instagram accounts</a> were fake. According to The New York Times, there's a cottage industry of companies which will automatically like, follow, and comment on Instagram on behalf of their client accounts.</p><p>The company <a href="http://uk.businessinsider.com/instagram-cracks-down-on-fake-accounts-by-shutting-down-instagress-2017-4">began seriously cracking down on bots a year ago</a> when it closed down Instagress, a popular automated like and comment service.</p><p>Last year, <a href="http://fortune.com/2017/06/29/instagram-artificial-intelligence-offensive-comments/">Instagram also debuted AI tools designed to fight trolls and spam</a>. It's likely that Instagram is using the vast trove of user behaviour data that it collects to train its AI to recognise real humans versus bot-controlled accounts. The end result is likely akin to a score of how likely an account is to be a human or a bot, and Instagram will act accordingly.</p><p>Business Insider has contacted Instagram for clarification.</p>    <div class="margin-top">        <h3>NOW WATCH: <a href="http://www.businessinsider.com/which-macbook-buying-guide-2018-5">How to know which MacBook you should buy</a></h3>        <div>            <iframe class="bijwplayerclass" data-player="jwplayer" src="https://content.jwplatform.com/players/MY66B8lK-CEZHWCBG.html" width="100%" height="100%" frameborder="0" scrolling="auto" style="margin:0px;" allowfullscreen=""></iframe><script>function receiveMessage(event){ console.log(event.data.fullscreen); Android.fullscreen(event.data.fullscreen); }window.addEventListener("message", receiveMessage, false);</script>        </div>    </div>

Instagram says it tracks how you tap and scroll inside the app

<p><div class="KonaFilter image-container display-table"><div><div class="image on-image"><img src="https://static-ssl.businessinsider.com/image/5af9e86efe3eea18008b492b-1893/gettyimages-944769724.jpg" border="0" alt="facebook ceo mark zuckerberg" data-mce-source="Chip Somodevilla/Getty Images" data-mce-caption="Facebook co-founder, Chairman and CEO Mark Zuckerberg testifies before the House Energy and Commerce Committee in the Rayburn House Office Building on Capitol Hill April 11, 2018 in Washington, DC." /><p class="source"><span>Chip Somodevilla/Getty Images</span></p><p class="caption">Facebook co-founder, Chairman and CEO Mark Zuckerberg testifies before the House Energy and Commerce Committee in the Rayburn House Office Building on Capitol Hill April 11, 2018 in Washington, DC.</p></div><p></p></div></div></p><p>Facebook banned 583 million fake accounts in the first three months of 2018, the social network has revealed. It's a major effort towards transparency from Facebook in the wake of the Cambridge Analytica scandal.</p><p>The Silicon Valley social networking giant announced on Tuesday that it will publish data about the scale and effectiveness of its content moderation efforts - offering outsiders an unprecedented look at Facebook's efforts to keep its platform free from objectionable, rule-breaking, and sometimes illegal content.</p><p>Here are some of the key numbers for the first three months of 2018: </p><ul><li><strong>Facebook "took action" on 3.4 million pieces of content that contained graphic violence</strong>. In this particular case, the "action" consisted of placing a warning over the content rather than removing it - at least, in the "vast majority" of cases. Of all the content viewed by users, between 0.22-0.27% of views was of rule-breaking content showing graphic violence. Which is to say, this doesn't mean that 0.22% of the content posted on Facebook contained graphic violence; just that the graphic content posted accounted for 0.22% of total views.</li><li><strong>It took action on 21 million pieces of content containing nudity and sexual activity</strong>. 0.07-0.09% of views were of the content.</li><li><strong>Facebook took action on 1.9 million pieces of content over terrorist propaganda</strong>. The company didn't provide a number of views, but said it was "extremely low."</li><li><strong>Facebook took action on 2.5 million pieces of content over hate speech</strong>, but doesn't have view numbers as it is still "developing measurement methods for this violation type."</li><li><strong>It took action on 837 million pieces of spam</strong>, though it didn't have view numbers.</li><li><strong>It disabled 583 million fake accounts</strong>. It estimates that overall, 3-4% of its monthly active users are fake.</li></ul><p>The data also illustrates where Facebook's AI moderation systems are effectively identifying and taking down problematic content - and the areas where it still struggles to identify problems. 99.5% of terrorist content was detected and taken down by Facebook before users reported it, for example, while just 38% of hate speech was detected without user-submitted reports.</p><p>The company plans to release the data periodically in the future, the<a href="https://transparency.facebook.com"> same way it already releases reports on the numbers of government data requests and intellectual property takedown requests it receives</a>.</p><p>The new disclosures come <a href="http://uk.businessinsider.com/facebook-publishes-community-standards-moderators-2018-4">after Facebook published for the first time its full "Community Standards"</a>: The internal rules it uses to determine exactly what is and isn't allowed on the social network. The full guidelines are about 8,500 words long, and and go into explicit detail around what is and isn't acceptable in terms of violent, sexual, or otherwise controversial content, along with hate speech and threatening language.</p><p>These releases come in the wake of the Cambridge Analytica scandal, which has left the company battling to restore its reputation with users and developers - though employees have said the decision to release the Community Standards was not driven by recent events.</p>    <div class="margin-top">        <h3>NOW WATCH: <a href="http://www.businessinsider.com/google-duplex-home-assistant-full-conversations-make-appointments-reservations-2018-5">Google's new AI can impersonate a human to schedule appointments and make reservations</a></h3>        <div>            <iframe class="bijwplayerclass" data-player="jwplayer" src="https://content.jwplatform.com/players/9NaapOJm-CEZHWCBG.html" width="100%" height="100%" frameborder="0" scrolling="auto" style="margin:0px;" allowfullscreen=""></iframe><script>function receiveMessage(event){ console.log(event.data.fullscreen); Android.fullscreen(event.data.fullscreen); }window.addEventListener("message", receiveMessage, false);</script>        </div>    </div>

Facebook banned 583 million fake accounts over 3 months, in an unprecedented display of transparency

How diehard Trump fans transformed their Twitter accounts into bots which spread conspiracies in a vast Russia-style disinformation network

<img src="https://static-ssl.businessinsider.com/image/5bd768c6d1e844043a215df2-2371/gettyimages-126695011.jpg" alt="Jeff weiner" /><p class="source"><span>Stephen Lam/Getty Images</span></p><p class="caption">LinkedIn CEO Jeff Weiner</p><ul class="summary-list"><li>Three years after Microsoft bought it for $26.2 billion, LinkedIn is making the move to the Microsoft Azure cloud.</li><li>The move was first announced in July 2019, but the company pegs the project's completion date within the next three and a half years, LinkedIn CTO and VP of Engineering Raghu Hiremagalur estimated to Business Insider.</li><li>Hiremagalur says the idea to move from its own data centers to Azure came from LinkedIn itself, and not Microsoft. </li><li>The appeal of Azure is the opportunity to take advantage of Microsoft's global footprint, storage and compute power, and security infrastructure to open new doors for LinkedIn.</li><li><a href="https://www.businessinsider.com/prime">Click here to read more BI Prime stories</a>.</li></ul><p>Last July, professional social network LinkedIn announced that it will move completely to the Microsoft Azure cloud.</p><p>This may seem overdue. LinkedIn has already been <a href="https://www.businessinsider.com/microsoft-buys-linkedin-2016-6">under Microsoft's umbrella</a> for about three years, following a blockbuster $26.2 billion acquisition.</p><p>However, LinkedIn has from the very beginning hosted itself from its own private data centers: <a href="https://www.businessinsider.com/linkedin-may-move-its-10000-employees-off-google-apps-2017-1/">Microsoft gave LinkedIn an unusual amount of control</a> over its fate after the deal closed, and made no apparent move to force any changes to its technology infrastructure.</p><p>But now, LinkedIn will move to Azure because of the "opportunity, not need," Raghu Hiremagalur, LinkedIn CTO and vice president of engineering, tells Business Insider. This move will take about three and a half years, Hiremagalur estimates.</p><p>However, he says it's worth the time and hassle: LinkedIn works just fine as it runs today, but moving to the cloud gives it more opportunities to take advantage of future technologies, he said. He also highlights Azure's major presence in data centers all over the globe, which he says will help LinkedIn better serve customers around the world. </p><p>"We made this decision primarily to significantly increase our agility in our infrastructure," Hiremagalur told Business Insider. "From what we see, the way innovation has happened in public cloud, specifically in Azure, we would like access to such technologies moving forward."</p><p>"Overall at the high level, we are doing this because we see this as an opportunity for us right now, not because we have a need to do this," Hiremagalur said. "There's a very deep sense of commitment from the Azure leadership team and Linkedin to make this a success," he later added.</p><h2>'Obviously, they were happy we were interested'</h2><p>Although Microsoft owns LinkedIn, it didn't ask the networking site to move to its cloud, Hiremagalur says. Instead, LinkedIn initiated the conversation. That said, he says, they didn't even think about moving to rival clouds like Google Cloud or Amazon Web Services.</p><p>"We reached out to folks at Microsoft and said we believe this is the right next step in the long term for us," Hiremagalur said. "Obviously, they were happy we were interested."</p><p>The cloud is attractive to LinkedIn because it can help manage its computing and storage needs more efficiently - public cloud platforms like Azure allow customers to grow or shrink their infrastructure depending on the capacity needed in the moment.</p><p>Hiremagalur also praises Azure's security, which uses advanced AI and other cutting-edge techniques, reducing LinkedIn's reliance on more traditional firewalls.</p><h2>Slow and steady</h2><p>LinkedIn plans to move onto Azure one region at a time. Hiremagalur estimates that in 2021, LinkedIn will be ready to start moving customer-facing traffic to Azure in at least one such region. They're taking it slow and steady, he said, as he doesn't want there to be any kind of disruption for users between now and the completion of the project.</p><p>The ultimate goal is that by the time the project is finished, LinkedIn will be completely off its data centers and entirely in the cloud. The company did explore the possibility of using hybrid cloud technology to connect, rather than replace, its data centers with Azure, but  the team decided against it because it "creates more complexity."</p><p>"Homogeneity in software stack overall is a big plus," Hiremagalur said. "That's why hybrid is not something we want."</p><p>The team isn't exactly starting from scratch, though. LinkedIn already has a tight collaboration with the Azure team, Hiremagalur says, and already uses Azure's capabilities for spam detection. It's starting to explore the possibility of building Azure-powered machine learning features, too. This is just the start, though, he says.</p><p>"I would like to over time, figure out ways to embrace more native Azure capabilities," Hiremagalur said. "I want us to start leaning on Azure for things they are very good at so we can start focusing on things we're good at."</p><p>By that same token, though, Hiremagalur said that it'll take some planning to get there.</p><p>"We want to leverage Azure for what we see as an opportunity - the kinds of networking backbone that Azure has, the investments they're making, and storage innovations," Hiremagalur said. "Evolutions of these things will become a reality. Those are things we want access to."</p><p><em><strong>Got a tip?</strong> Contact this reporter via email at <a href="mailto:rmchan@businessinsider.com">rmchan@businessinsider.com</a>, Signal at 646.376.6106, Telegram at @rosaliechan, or Twitter DM at <a href="https://twitter.com/rosaliechan17">@rosaliechan17</a>. (PR pitches by email only, please.) Other types of secure messaging available upon request. You can also <a href="https://www.businessinsider.com/how-to-tip-business-insider-securely-guide-signal-securedrop-2017-6">contact Business Insider securely via SecureDrop</a>.</em></p><p class="tagline"><a href="https://www.businessinsider.com/intelligence/bi-intelligence-daily?IR=T&amp;itm_source=businessinsider&amp;itm_medium=ctgr_bii&amp;itm_term=ctgr_subscription&amp;itm_content=microsoft&amp;itm_campaign=bi-intelligence-daily"><img src="https://static.businessinsider.com/image/5dc428a57eece546b8629657" alt="Signup Today: Free Daily Newsletter from Business Insider Intelligence" style="width: 100% !important" /></a></p>            <h3>NOW WATCH: <a href="https://www.businessinsider.com/stop-using-champagne-flutes-this-is-the-best-way-to-drink-champagne-2017-12">Stop using champagne flutes - this is the best way to drink champagne</a></h3>                    <iframe class="bijwplayerclass" src="https://content.jwplatform.com/players/JCKIRytm-CEZHWCBG.html" width="100%" height="100%" frameborder="0" scrolling="auto" style="margin:0px;"></iframe>            

You Searched For "AI spam detection" and got 9 results

From 'harass dash' to 'dark social,' 25 Facebook code words revealed in leaked whistleblower documents - and what they all mean

Hannah Towey

WhatsApp banned two million Indian accounts during May 15-June 15 period

PTI

Android malware 'BlackRock' prowling in cyber space, may steal banking data: Advisory

PTI

Responsible use of artificial intelligence can make our financial markets even stronger

Kirsten Wegner, Contributor

Instagram says it tracks how you tap and scroll inside the app

Shona Ghosh

Facebook banned 583 million fake accounts over 3 months, in an unprecedented display of transparency

Rob Price

How diehard Trump fans transformed their Twitter accounts into bots which spread conspiracies in a vast Russia-style disinformation network

Tom Porter,Tom Porter

The CTO of LinkedIn explains how and why it's making the massive shift to Microsoft's cloud, three years after the $26.2 billion acquisition

Rosalie Chan

35. You've got Enron mail!

Dan Bobkoff,Amy Pedulla,Jenni…