{"id":206240,"date":"2024-02-17T08:50:56","date_gmt":"2024-02-17T08:50:56","guid":{"rendered":"https:\/\/michigandigitalnews.com\/index.php\/2024\/02\/17\/realism-of-openais-sora-video-generator-raises-security-concerns\/"},"modified":"2025-06-25T17:21:49","modified_gmt":"2025-06-25T17:21:49","slug":"realism-of-openais-sora-video-generator-raises-security-concerns","status":"publish","type":"post","link":"https:\/\/michigandigitalnews.com\/index.php\/2024\/02\/17\/realism-of-openais-sora-video-generator-raises-security-concerns\/","title":{"rendered":"Realism of OpenAI\u2019s Sora video generator raises security concerns"},"content":{"rendered":"<p> [ad_1]<br \/>\n<\/p>\n<div id=\"\">\n<figure class=\"article-image-inline ArticleImage\" data-method=\"caption-shortcode\">\n<div class=\"ArticleImage__Wrapper\"><img fetchpriority=\"high\" decoding=\"async\" src=\"https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=1200\" srcset=\"https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=100 100w, https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=200 200w, https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=249 249w, https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=300 300w, https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=400 400w, https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=500 500w, https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=600 600w, https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=700 700w, https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=800 800w, https:\/\/images.newscientist.com\/wp-content\/uploads\/2024\/02\/16201946\/SEI_191815757.jpg?width=900 900w\" class=\"image size-full wp-image-2417647 ReplaceImageLazyload\" sizes=\"(min-width: 1130px) 900px, (min-width: 1025px) 900, (min-width: 768px) calc(100vw - 30px), calc(100vw - 30px)\" alt=\"\" width=\"1350\" height=\"899\" data-credit=\"Sora\/OpenAI\" data-caption=\"The AI program Sora generated a video featuring this artificial woman based on a text prompt\"\/><\/div><figcaption class=\"ArticleImageCaption\">\n<div class=\"ArticleImageCaption__CaptionWrapper\">\n<p class=\"ArticleImageCaption__Title\">The AI program Sora generated a video featuring this artificial woman based on a text prompt<\/p>\n<p class=\"ArticleImageCaption__Credit\">Sora\/OpenAI<\/p>\n<\/div>\n<\/figcaption><\/figure>\n<\/p>\n<p>OpenAI has unveiled its latest artificial intelligence system, a program called Sora that can transform text descriptions into photorealistic videos. The video generation model is spurring excitement about advancing AI technology, along with growing concerns over how artificial deepfake videos worsen misinformation and disinformation during a pivotal election year worldwide.<\/p>\n<p>\n    <iframe title=\"OpenAI&#039;s &#039;Sora&#039; can generate realistic videos with just a single prompt\" width=\"1200\" height=\"675\" src=\"https:\/\/www.youtube.com\/embed\/MfL-Kalpozc?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share\" allowfullscreen><\/iframe><br \/>\n    <\/iframe>\n<\/p>\n<p>The Sora AI model can currently create videos up to 60 seconds long using either text instructions alone or text combined with an image. One demonstration video starts with a text prompt that describes how \u201ca stylish woman walks down a Tokyo street filled with warm glowing neon and animated city signage\u201d. Other examples include a dog frolicking in the snow, vehicles driving along roads and more fantastical scenarios such as sharks swimming in midair between city skyscrapers.<\/p>\n<p>\u201cAs with other techniques in generative AI, there is no reason to believe that text-to-video will not continue to rapidly improve \u2013 moving us closer and closer to a time when it will be difficult to distinguish the fake from the real,\u201d says <a href=\"https:\/\/www.ischool.berkeley.edu\/people\/hany-farid\">Hany Farid<\/a> at the University of California, Berkeley. \u201cThis technology, if combined with AI-powered voice cloning, could open up an entirely new front when it comes to creating deepfakes of people saying and doing things they never did.\u201d<\/p>\n<p><span class=\"js-content-prompt-opportunity\"\/><\/p>\n<p>Sora is based in part on OpenAI\u2019s preexisting technologies, such as the image generator DALL-E and the GPT large language models. Text-to-video AI models have lagged somewhat behind those other technologies in terms of realism and accessibility, but the Sora demonstration is an \u201corder of magnitude more believable and less cartoonish\u201d than what has come before, says <a href=\"https:\/\/www.linkedin.com\/in\/racheltobac\/\">Rachel Tobac<\/a>, co-founder of SocialProof Security, a white-hat hacking organisation focused on social engineering.<\/p>\n<p>To achieve this higher level of realism, Sora combines two different AI approaches. The first is a diffusion model similar to those used in AI image generators such as DALL-E. These models learn to gradually convert randomised image pixels into a coherent image. The second AI technique is called \u201ctransformer architecture\u201d and is used to contextualise and piece together sequential data. For example, large language models use transformer architecture to assemble words into generally comprehensible sentences. In this case, OpenAI broke down video clips into visual \u201cspacetime patches\u201d that Sora\u2019s transformer architecture could process.<\/p>\n<p>Sora\u2019s videos still contain plenty of mistakes, such as a walking human\u2019s left and right legs swapping places, a chair randomly floating in midair or a bitten cookie magically having no bite mark. Still, <a href=\"https:\/\/twitter.com\/DrJimFan\/status\/1758210245799920123?t=8vuFpQy8YSXh6J4tW32iqQ&amp;s=19\">Jim Fan<\/a>, a senior research scientist at NVIDIA, took to the social media platform X to praise Sora as a \u201cdata-driven physics engine\u201d that can simulate worlds.<\/p>\n<p>The fact that Sora\u2019s videos still display some strange glitches when depicting complex scenes with lots of movement suggests that such deepfake videos will be detectable for now, says <a href=\"https:\/\/www.cs.princeton.edu\/~arvindn\/\">Arvind Narayanan<\/a> at Princeton University. But he also cautioned that in the long run \u201cwe will need to find other ways to adapt as a society\u201d.<\/p>\n<p>OpenAI has held off on making Sora publicly available while it performs \u201cred team\u201d exercises where experts try to break the AI model\u2019s safeguards in order to assess its potential for misuse. The select group of people currently testing Sora are \u201cdomain experts in areas like misinformation, hateful content and bias\u201d, says an OpenAI spokesperson.<\/p>\n<p>This testing is vital because artificial videos could let bad actors generate false footage in order to, for instance, harass someone or sway a political election. Misinformation and disinformation fuelled by AI-generated deepfakes ranks as a major concern <a href=\"https:\/\/www.weforum.org\/publications\/global-risks-report-2024\/\">for leaders<\/a> in academia, business, government and other sectors, as well as <a href=\"https:\/\/aiimpacts.org\/wp-content\/uploads\/2023\/04\/Thousands_of_AI_authors_on_the_future_of_AI.pdf\">for AI experts<\/a>.<\/p>\n<p>\u201cSora is absolutely capable of creating videos that could trick everyday folks,\u201d says Tobac. \u201cVideo does not need to be perfect to be believable as many people still don\u2019t realise that video can be manipulated as easily as pictures.\u201d<\/p>\n<p>AI companies will need to collaborate with social media networks and governments to handle the scale of misinformation and disinformation likely to occur once Sora becomes open to the public, says Tobac. Defences could include implementing unique identifiers, or \u201cwatermarks\u201d, for AI-generated content.<\/p>\n<p>When asked if OpenAI has any plans to make Sora more widely available in 2024, the OpenAI spokesperson described the company as \u201ctaking several important safety steps ahead of making Sora available in OpenAI\u2019s products\u201d. For instance, the company already uses automated processes aimed at preventing its commercial AI models from generating depictions of extreme violence, sexual content, hateful imagery and real politicians or celebrities. With more people than ever before <a href=\"https:\/\/apnews.com\/article\/global-elections-2024-preview-cb77b0940964c5c95a9affc8ebb6f0b7\">participating in elections this year<\/a>, those safety steps will be crucial.<\/p>\n<section class=\"ArticleTopics\">\n<p class=\"ArticleTopics__Heading\">Topics:<\/p>\n<ul class=\"ArticleTopics__List\">\n<li class=\"ArticleTopics__ListItem\"><a class=\"ArticleTopics__ListItemLink\" href=\"https:\/\/www.newscientist.com\/article-topic\/artificial-intelligence\/\" data-analytics-hook=\"topics-link\">artificial intelligence<\/a><span>\/<\/span><\/li>\n<li class=\"ArticleTopics__ListItem\"><a class=\"ArticleTopics__ListItemLink\" href=\"https:\/\/www.newscientist.com\/article-topic\/video\/\" data-analytics-hook=\"topics-link\">video<\/a><\/li>\n<\/ul>\n<\/section><\/div>\n<p><script async src=\"\/\/platform.twitter.com\/widgets.js\" charset=\"utf-8\"><\/script><br \/>\n<br \/>[ad_2]<br \/>\n<br \/><a href=\"https:\/\/www.newscientist.com\/article\/2417639-realism-of-openais-sora-video-generator-raises-security-concerns\/?utm_campaign=RSS%7CNSNS&#038;utm_source=NSNS&#038;utm_medium=RSS&#038;utm_content=home\">Source link <\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>[ad_1] The AI program Sora generated a video featuring this artificial woman based on a text prompt Sora\/OpenAI OpenAI has unveiled its latest artificial intelligence<\/p>\n","protected":false},"author":1,"featured_media":206241,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"_uf_show_specific_survey":0,"_uf_disable_surveys":false,"footnotes":""},"categories":[177],"tags":[],"_links":{"self":[{"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/posts\/206240"}],"collection":[{"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/comments?post=206240"}],"version-history":[{"count":3,"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/posts\/206240\/revisions"}],"predecessor-version":[{"id":343745,"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/posts\/206240\/revisions\/343745"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/media\/206241"}],"wp:attachment":[{"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/media?parent=206240"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/categories?post=206240"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/michigandigitalnews.com\/index.php\/wp-json\/wp\/v2\/tags?post=206240"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}