{"id":66948,"date":"2023-04-13T08:13:00","date_gmt":"2023-04-13T07:13:00","guid":{"rendered":"https:\/\/mediamakersmeet.com\/?p=66948"},"modified":"2023-04-13T09:16:57","modified_gmt":"2023-04-13T08:16:57","slug":"how-to-counter-ai-generated-disinformation","status":"publish","type":"post","link":"https:\/\/mediamakersmeet.com\/how-to-counter-ai-generated-disinformation\/","title":{"rendered":"How to counter AI-generated disinformation"},"content":{"rendered":"\n<p class=\"has-background\" style=\"background-color:#fffeee\">Tech journalists can tell us\u00a0<em>how<\/em>\u00a0AI generates disinformation. We need specialist internet culture reporters to tell us\u00a0<em>why<\/em>.<\/p>\n\n\n\n<p>The pope looks great in his big jacket, eh? The image of God\u2019s earthly representative showing off his drip was all over the internet recently, shared with a mix of amusement and horror of what it represents. <\/p>\n\n\n\n<div style=\"height:15px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<figure class=\"wp-block-image size-full\"><img fetchpriority=\"high\" decoding=\"async\" width=\"959\" height=\"475\" src=\"https:\/\/mediamakersmeet.com\/wp-content\/uploads\/2023\/04\/image-6.png\" alt=\"\" class=\"wp-image-66950\" srcset=\"https:\/\/mediamakersmeet.com\/wp-content\/uploads\/2023\/04\/image-6.png 959w, https:\/\/mediamakersmeet.com\/wp-content\/uploads\/2023\/04\/image-6-400x198.png 400w, https:\/\/mediamakersmeet.com\/wp-content\/uploads\/2023\/04\/image-6-768x380.png 768w\" sizes=\"(max-width: 959px) 100vw, 959px\" \/><\/figure>\n\n\n\n<div style=\"height:15px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p>The AI-generated image \u2013 one of many thousands currently being generated at scale \u2013 looks realistic enough to fool all but the most careful of observers, and that\u2019s where the big issue arises.<\/p>\n\n\n\n<figure class=\"wp-block-embed is-type-rich is-provider-twitter wp-block-embed-twitter\"><div class=\"wp-block-embed__wrapper\">\n<blockquote class=\"twitter-tweet\" data-width=\"550\" data-dnt=\"true\"><p lang=\"en\" dir=\"ltr\">We are completely and utterly unprepared for how to deal with this <a href=\"https:\/\/t.co\/yM0JUSMSFN\">https:\/\/t.co\/yM0JUSMSFN<\/a><\/p>&mdash; Esther Kezia Thorpe (@EstherKeziaT) <a href=\"https:\/\/twitter.com\/EstherKeziaT\/status\/1640273029270839298?ref_src=twsrc%5Etfw\">March 27, 2023<\/a><\/blockquote><script async src=\"https:\/\/platform.twitter.com\/widgets.js\" charset=\"utf-8\"><\/script>\n<\/div><\/figure>\n\n\n\n<p>Manipulated images are nothing new. We\u2019ve had panics about airbrushing, photoshop, deepfakes etc. for years now. What&nbsp;<em>is<\/em>&nbsp;different is the sheer ease and sophistication with which almost anyone can use generative AI to create images, synthesised voices and even video, at scale. The opportunities for the deliberate spreading of disinformation are obvious, and there are many tech reporters busily explaining&nbsp;<em>how<\/em>&nbsp;AI goes about its creation.<\/p>\n\n\n\n<p>Just as worrying, however, is the scale with which that creation is now possible: the balance of signal to noise is tipping wildly towards the latter. Public discourse is already flooded with vast amounts of misinformation, and a second deluge is on the way.<\/p>\n\n\n\n<p>Newspapers obviously have a role to play here. They, and fact-checking sites like Snopes, will prove ever more vital in the age of AI-generated disinformation. But to counter it, they need to invest in new roles \u2013 not just building out tech desks, but hiring internet culture reporters.<\/p>\n\n\n\n<p>Even if tech journalists help us understand&nbsp;<em>how<\/em>&nbsp;AI takes a prompt and creates audio of, say,&nbsp;<a href=\"https:\/\/youtu.be\/YtDLOq2JODo\" target=\"_blank\" rel=\"noreferrer noopener\">Biden, Obama and Trump ranking the&nbsp;<em>Dark Souls<\/em>&nbsp;series<\/a>, the public doesn\u2019t necessarily know&nbsp;<em>why<\/em>. In that specific case it\u2019s simple \u2013 it\u2019s funny, and that explanation suffices. But when the created image is something as semi-plausible as \u2018the Pope wears a big coat\u2019 or \u2018<a href=\"https:\/\/www.snopes.com\/fact-check\/trump-getting-arrested-ai-image\/\" target=\"_blank\" rel=\"noreferrer noopener\">Trump gets arrested<\/a>\u2019 then there\u2019s a need for the paper to discover the motivation behind its creation and dissemination.<\/p>\n\n\n\n<p>Sometimes that will be a serious investigation explaining why a state is flooding the internet with generated images designed to sow confusion about its activities. More often, knowing the internet, it will be something trivial like a generated image of two celebrities \u201ccaught\u201d in a tryst.&nbsp;<\/p>\n\n\n\n<p>Despite the wildly varying seriousness of those cases, internet culture reporters will be required to explain them to a public that does not have the time to trawl subreddits and message boards to understand the context. And even when the created image is trivial in its importance it will be incumbent upon newspapers to debunk them to build up trust in their capabilities, so that when something seismic happens the public knows where to turn to check if photos and videos of the event are real.\u00a0<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Context at scale<\/h3>\n\n\n\n<p>Internet culture is, to put it mildly,&nbsp;<em>wild<\/em>. There are subreddits and communities so steeped in in-jokes and internal histories that they are mostly impenetrable to outsiders. I\u2019m a member of r\/TwoBestFriendsPlay, for instance, which is a still-extant subreddit dedicated to a YouTube channel that closed down four years ago.&nbsp;<a href=\"https:\/\/i.redd.it\/mvxoafbgkv881.png\" target=\"_blank\" rel=\"noreferrer noopener\">This image<\/a>&nbsp;is perfectly understandable to me, but I\u2019m willing to bet nobody reading this will be able to parse it. I\u2019d also bet that you\u2019re also members of online communities that are similarly obtuse to newcomers.&nbsp;<\/p>\n\n\n\n<p>And unfortunately, a lot of the motivation for someone to use AI to generate an image or video is going to be buried in communities like those. It might not be&nbsp;<em>as<\/em>&nbsp;insular as the TBFP play example above, but we are going to&nbsp;need&nbsp;reporters who are used to parsing internet subcommunities to provide that context at speed.&nbsp;<\/p>\n\n\n\n<p>It\u2019s also important that those reporters stay abreast of the communities who are driving the tech forward. For the moment the r\/stablediffusion subreddit is mostly dedicated to, well, predictably,&nbsp;<em>this<\/em>\u2026<\/p>\n\n\n\n<blockquote class=\"reddit-embed-bq\" style=\"height:500px\" data-embed-height=\"500\">      <a href=\"https:\/\/www.reddit.com\/r\/StableDiffusion\/comments\/123dl2v\/the_absolute_state_of_sd_twitter_people_will\/\">The absolute state of SD twitter. People will start to have a very skewed view of AI generated content soon.<\/a><br> by      <a href=\"https:\/\/www.reddit.com\/user\/Great_Ape_17\">u\/Great_Ape_17<\/a> in      <a href=\"https:\/\/www.reddit.com\/r\/StableDiffusion\/\">StableDiffusion<\/a>    <\/blockquote><script async=\"\" src=\"https:\/\/embed.reddit.com\/widgets.js\" charset=\"UTF-8\"><\/script>\n\n\n\n<p>\u2026 but there are people on there who are pushing text-to-video forward, and who will be responsible for democratising the tech that enables anyone to create it.&nbsp;<\/p>\n\n\n\n<p>None of this is to say that publishers aren\u2019t already doing some of this. As I mentioned above there are many tech reporters who are grimly acknowledging Chat-GPT, Bard and generative AI will probably be their major beat for the next few years. The BBC recognised the threat already, with its dedicated misinformation reporter\u00a0<a href=\"https:\/\/voices.media\/bbcs-specialist-disinformation-reporter-marianna-spring-proactively-countering-conspiracies\/\">Marianna Spring<\/a>, and its just-announced BBC Bitesize service that aims to protect kids from this specific type of disinformation.<\/p>\n\n\n\n<p>BBC Creative\u2019s Executive Creative Director, Rasmus Smith Bech, says: \u201cAt the BBC we have a duty to protect kids from misinformation online. Don\u2019t Learn Off Randoms explores the crazy world of randomness on the internet, from unicorn-believers to Flat Earthers, offering up the trustworthy services of Bitesize as an antidote to the wild west of the wider internet.\u201d<\/p>\n\n\n\n<p>In fact Newsweek, Rolling Stone and many other titles already have internet culture correspondents. A lot of other titles aren\u2019t just debunking, they\u2019re&nbsp;<a href=\"https:\/\/www.poynter.org\/fact-checking\/2023\/tips-for-detecting-doctored-or-out-of-context-videos\/\" target=\"_blank\" rel=\"noreferrer noopener\">documenting how they did so<\/a>&nbsp;in the name of media literacy. We have had experts in navigating 4chan and other shitposting boards for years. That\u2019s all great, effective and laudable.<\/p>\n\n\n\n<p>But in the face of this new wave of disinformation, the industry needs\u00a0<em>so many more<\/em>\u00a0reporters, embedded in digital communities and reporting on not just how AI-generated art is created, but why. As people look for answers, the context is just as important as the content.<\/p>\n\n\n\n<div style=\"height:100px\" aria-hidden=\"true\" class=\"wp-block-spacer\"><\/div>\n\n\n\n<p><em>Republished with kind permission of\u00a0<a href=\"http:\/\/voices.media\/\">Media Voices<\/a>,\u00a0a weekly look at all the news and views from across the media world.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Tech journalists can tell us\u00a0how\u00a0AI generates disinformation. We need specialist internet culture reporters to tell us\u00a0why. The pope looks great in his big jacket, eh? The image of God\u2019s earthly representative showing off his drip was all over the internet recently, shared with a mix of amusement and horror of what it represents. The AI-generated [&hellip;]<\/p>\n","protected":false},"author":4,"featured_media":66954,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[17613,26],"tags":[],"yst_prominent_words":[1469,4342,6941,1735,9741,3114,10075,2452,4404,1947,1774,1330],"acf":[],"_links":{"self":[{"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/posts\/66948"}],"collection":[{"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/users\/4"}],"replies":[{"embeddable":true,"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/comments?post=66948"}],"version-history":[{"count":0,"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/posts\/66948\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/media\/66954"}],"wp:attachment":[{"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/media?parent=66948"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/categories?post=66948"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/tags?post=66948"},{"taxonomy":"yst_prominent_words","embeddable":true,"href":"https:\/\/mediamakersmeet.com\/wp-json\/wp\/v2\/yst_prominent_words?post=66948"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}