Latest news with #Geoconfirmed
BBC News
3 hours ago
- Politics
- BBC News
AI disinformation wey dey comot from di Iran- Israel War
Plenti disinformation don comot online since wen Israel start to dey strike Iran last week, dis na as BBC Verify don review plenti posts for social media wey dem use to boost how effective di Tehran response dey. Di analysis find out some videos wey dem use artificial intelligence to make, wey bin dey brag about Iran military capacity togeda wit fake clips wey show wetin strikes cause for Israeli targets. Di three most viewed fake videos wey BBC Verify find don get ova 100 million views from all di platforms wey dem follow post dem. Pro-Israeli accounts sef don also share dia own disinformation online. Dis na mainly by say dem repost old clips of protests and gatherings for Iran wey dem go lie say dem dey show say di public for Iran no dey feel dia goment and dey support di Israel military campaign. Israel bin launch strikes for Iran for 13 June wey cause Iran to send several rounds of missile and drone attacks to Israel. One joinbodi wey dia work na to torchlight and analyse open source images tok say di amount of disinformation wey dey online dey, "astonishing" and even accuse who dem call "engagement farmers" say dem dey try to make money from di kasala by sharing misleading content wey dey made to get attention online. Geoconfirmed, wey be di online verification group write for X say, "we dey see everting from unrelated videos from Pakistan, to recycled videos from di October 2024 strikes, wey don gada ova 20 million views, to game videos and AI generated content wey dem dey try pass off say na wetin happun be dat". Some accounts don turn "super spreaders" of disinformation as dem dey get growth for dia followers. One pro Iranian account wey no look like e get any ties wit Tehran don grow dia followers from ova 700,000 for 13 June to 1.4 million for 19 June. Dis na 100% growth and, e neva even reach one week sef. Na one of di many not known accounts wey don dey show for pipo feeds as of recent. All of dem get blue ticks, get ogbonge messaging and don dey post disinformation well-well. Bicos some dey use names wey look like say dem dey official, so pipo go reason say dem bi real accounts but e no dey clear who dey run di accounts. Dis amount of disinformation dey marked as "di first time wey we go see di use of generative AI for dis level during kasala", na wetin Emmanuella Saliba wey be di Chief Investigative Officer for di analyst group Get Real tell BBC Verify. Accounts wey BBC Verify review bin dey normally show AI generated pictures as e be like say dem wan exaggerate how di Iran response to Israel strikes dey successful. One image wey get 27m views show dozens of missiles as dem dey enta di city of Tel Aviv. Anoda video look like e show missile strike on top one building for di Israeli city late for night. Ms Saliba say dis clips dey normally show night time attacks wey dey make am harder to verify. Di AI fakes dey also focus on claims say dem destroy di Israeli F-35 fighter jets, wey be di state of di art planes US make wey fit strike both targets for ground and for air. If di plenti clips wey dey comot dey real, dat mean say Iran for don destroy 15% of di Israel fleet of fighters. Dis na wetin Lisa Kaplan wey be di CEO of di Alethea analyst group tell BBC Verify. We neva fit verify any video of F-35s wey dem shoot down. One of di post wey don go far claim to show jet wey dey damaged afta dem shoot am down for Iranian desert. But signs of AI manipulation bin dey inside, like how di pipo wey dey around di jets bin dey di same size as di nearby motor and how di sand no even move small to show say jet crash land on am. Anoda video wey get 21.1 million views for TikTok show one Israeli F-35 wey air defences bin shoot down, but wia di video bin actually come from na from one flight simulator video game. TikTok don comot di video afta BBC Verify meet wit dem. Ms Kaplan explain say some of di focus wey dey on top F-35s na from network of accounts wey Alethea bin formerly link to Russian influence operations. She bin note say Russian influence operations don shift from dia work to undermine support for di war for Ukraine to come sow doubts on top weda Western and especially American weapons dey work well. Ms Kaplan say, "Russia no really get response to di F-35. So wetin dem go do? Na to reduce di support e dey get for certain kontris." Disinformation also dey spread by well-known accounts wey don chook mouth for di Israel Gaza war and oda kasala bifor. Dia motivations dey different but sabi pipo tok say some of dem fit dey try to make money from di kasala as some social media platforms don dey offer money to accounts wey dey get plenti views. Pro-Israeli posts, howeva, dey focus on suggestions say di goment of Iran dey face backlash as di strikes dey go on. Among dem, na widely shared AI video wey lie, wey show Iranians dey hala "we love Israel" for di streets of Tehran. But recently, amd as di speculation about US strikes for Iran nuclear sites still dey grow, some accounts don start to dey post AI pictures of B-2 bombers ova Tehran. Pipo don dey torchlight di B-2 since Israel strikes for Iran start bicos na di only aircraft wey fit effectively carry out attack for Iran underground nuclear sites. Official sources for Iran and Israel don share some of di fake images. State media for Tehran share fake videos of strikes and AI image of crashed F-35 while Israel Defense Forces (IDF) gbab community note for X for one of dia post say dem use old unrelated footage of missile bombardment. Plenti of di disinformation wey BBC Verify review don dey shared for X. Sotay users now dey ask di platform AI chatbot, Grok weda di posts na true or na lie. But for some cases, Grok don tok say di AI videos dey real. One of dat kain video show plenti trucks wey dey carry ballistic missiles dey comot from one mountainside complex. Di signs of AI content show as di rocks for di video bin dey move dia self na wetin Ms Saliba tok. But Grok still tell X users say di videos bin dey real and also cite reports wey media outlets like Newsweek and Reuters put out say make pipo "check trusted news for clarity". X no ansa BBC Verify wen dem ask for comments for wetin Grok dey do. Many videos also comot for TikTok and Instagram, For statement to BBC Verify, TikTok tok say dem dey ginger to enforce community guidelines "wey prohibit inaccurate, misleading or fake content" and dem dey work wit independent fact checkers to "verify misleading content". Instagram owner Meta no ansa dia own request for comment. While motivation for pipo wey dey create dis online fakes dey plenti, many of dem dey shared by ordinary social media users. Matthew Facciani, wey be researcher for di University of Notre Dame, bin suggest say disinformation fit spread sharparly online wen pipo dey faced wit two options like di kain wey kasala and politics dey bring comot. E say, "dat one torchlight wetin be di bigger social and psychological issue of pipo wey wan reshare tins if e go wit dia political identity and also in general, more sensationalist emotional content wey dey spread quicker for di online world".
BBC News
14 hours ago
- Politics
- BBC News
Israel-Iran conflict unleashes wave of AI disinformation
A wave of disinformation has been unleashed online since Israel began strikes on Iran last week, with dozens of posts reviewed by BBC Verify seeking to amplify the effectiveness of Tehran's analysis found a number of videos - created using artificial intelligence - boasting of Iran's military capabilities, alongside fake clips showing the aftermath of strikes on Israeli targets. The three most viewed fake videos BBC Verify found have collectively amassed over 100 million views across multiple accounts have also shared disinformation online, mainly by recirculating old clips of protests and gatherings in Iran, falsely claiming that they show mounting dissent against the government and support among Iranians for Israel's military launched strikes in Iran on 13 June, leading to several rounds of Iranian missile and drone attacks on Israel. One organisation that analyses open-source imagery described the volume of disinformation online as "astonishing" and accused some "engagement farmers" of seeking to profit from the conflict by sharing misleading content designed to attract attention online."We are seeing everything from unrelated footage from Pakistan, to recycled videos from the October 2024 strikes—some of which have amassed over 20 million views—as well as game clips and AI-generated content being passed off as real events," Geoconfirmed, the online verification group, wrote on accounts have become "super-spreaders" of disinformation, being rewarded with significant growth in their follower count. One pro-Iranian account with no obvious ties to authorities in Tehran - Daily Iran Military - has seen its followers on X grow from just over 700,000 on 13 June to 1.4m by 19 June, an 85% increase in under a is one many obscure accounts that have appeared in people's feeds recently. All have blue ticks, are prolific in messaging and have repeatedly posted disinformation. Because some use seemingly official names, some people have assumed they are authentic accounts, but it is unclear who is actually running the torrent of disinformation marked "the first time we've seen generative AI be used at scale during a conflict," Emmanuelle Saliba, Chief Investigative Officer with the analyst group Get Real, told BBC secretive nuclear site that only a US bomb could hitUS moves 30 jets as Iran attack speculation growsIs the UK about to get dragged into Iran-Israel conflict?Accounts reviewed by BBC Verify frequently shared AI-generated images that appear to be seeking to exaggerate the success of Iran's response to Israel's strikes. One image, which has 27m views, depicted dozens of missiles falling on the city of Tel video purported to show a missile strike on a building in the Israeli city late at night. Ms Saliba said the clips often depict night-time attacks, making them especially difficult to fakes have also focussed on claims of destruction of Israeli F-35 fighter jets, a state-of-the art US-made plane capable of striking ground and air targets. If the barrage of clips were real Iran would have destroyed 15% of Israel's fleet of the fighters, Lisa Kaplan, CEO of the Alethea analyst group, told BBC Verify. We have yet to authenticate any footage of F-35s being shot widely shared post claimed to show a jet damaged after being shot down in the Iranian desert. However, signs of AI manipulation were evident: civilians around the jet were the same size as nearby vehicles, and the sand showed no signs of impact. Another video with 21.1 million views on TikTok claimed to show an Israeli F-35 being shot down by air defences, but the footage actually came from a flight simulator video game. TikTok removed the footage after being approached by BBC Verify. Ms Kaplan said that some of the focus on F-35s was being driven by a network of accounts that Alethea has previously linked to Russian influence noted that Russian influence operations have recently shifted course from trying to undermine support for the war in Ukraine to sowing doubts about the capability of Western - especially American - weaponry."Russia doesn't really have a response to the F-35. So what it can it do? It can seek to undermine support for it within certain countries," Ms Kaplan is also being spread by well-known accounts that have previously weighed in on the Israel-Gaza war and other motivations vary, but experts said some may be attempting to monetise the conflict, with some major social media platforms offering pay-outs to accounts achieving large numbers of contrast, pro-Israeli posts have largely focussed on suggestions that the Iranian government is facing mounting dissent as the strikes continuerAmong them is a widely shared AI-generated video falsely purporting to show Iranians chant "we love Israel" on the streets of in recent days - and as speculation about US strikes on Iranian nuclear sites grows - some accounts have started to post AI-generated images of B-2 bombers over Tehran. The B-2 has attracted close attention since Israel's strikes on Iran started, because it is the only aircraft capable of effectively carrying out an attack on Iran's subterranean nuclear sources in Iran and Israel have shared some of the fake images. State media in Tehran has shared fake footage of strikes and an AI-generated image of a downed F-35 jet, while a post shared by the Israel Defense Forces (IDF) received a community note on X for using old, unrelated footage of missile barrages.A lot of the Disinformation reviewed by BBC Verify has been shared on X, with users frequently turning to the platform's AI chatbot - Grok - to establish posts' veracity. However, in some cases Grok insisted that the AI videos were real. One such video showed an endless stream of trucks carrying ballistic missiles emerging from a mountainside complex. Tell-tale signs of AI content included rocks in the video moving of their own accord, Ms Saliba said. But in response to X users, Grok insisted repeatedly that the video was real and cited reports by media outlets including Newsweek and Reuters. "Check trusted news for clarity," the chatbot concluded in several messages.X did not respond to a request from BBC Verify for comment on the Chatbot's videos have also appeared on TikTok and Instagram. In a statement to BBC Verify, TikTok said it proactively enforces community guidelines "which prohibit inaccurate, misleading, or false content" and that it works with independent fact checkers to "verify misleading content". Instagram owner Meta did not respond to a request for comment. While the motivations of those creating online fakes vary, many are shared by ordinary social media Facciani, a researcher at the University of Notre Dame, suggested that disinformation can spread more quickly online when people are faced with binary choices, such as those raised by conflict and politics."That speaks to the broader social and psychological issue of people wanting to re-share things if it aligns with their political identity, and also just in general, more sensationalist emotional content will spread more quickly online." What do you want BBC Verify to investigate?
