{"id":219150,"date":"2025-10-12T12:39:07","date_gmt":"2025-10-12T12:39:07","guid":{"rendered":"https:\/\/www.newsbeep.com\/us\/219150\/"},"modified":"2025-10-12T12:39:07","modified_gmt":"2025-10-12T12:39:07","slug":"are-deepfakes-of-dead-people-rewriting-the-past","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/us\/219150\/","title":{"rendered":"Are deepfakes of dead people rewriting the past?"},"content":{"rendered":"

OpenAI\u2019s new text-to-video app, Sora, was supposed to be a social AI playground, allowing users to create imaginative AI videos of themselves, friends and celebrities while building off of others\u2019 ideas<\/a>.<\/p>\n

The social structure of the app, which allows users to adjust the availability of their likeness in others\u2019 videos, seemed to address the most pressing questions of consent around AI-generated video when it was launched last week. <\/p>\n

But as Sora sits atop<\/a> the iOS App Store<\/a> with over 1 million downloads<\/a>, experts worry about its potential to deluge the internet with historical misinformation and deepfakes of deceased historical figures who cannot consent to or opt out of Sora\u2019s AI models.<\/p>\n

In less than a minute, the app can generate short videos of deceased celebrities in situations they were never in: Aretha Franklin making soy candles, Carrie Fisher trying to balance on a slackline, Nat King Cole ice skating in Havana and Marilyn Monroe teaching Vietnamese to schoolchildren, for instance.<\/p>\n

That\u2019s a nightmare for people like Adam Streisand, an attorney who has represented several celebrity estates, including Monroe\u2019s at one point.<\/p>\n

\u201cThe challenge with AI is not the law,\u201d Streisand said in an email, pointing out that California\u2019s courts have long protected celebrities<\/a> \u201cfrom AI-like reproductions of their images or voices.\u201d<\/p>\n

\u201cThe question is whether a non-AI judicial process that depends on human beings will ever be able to play an almost 5th dimensional game of whack-a-mole.\u201d<\/p>\n

Videos on Sora range from the absurd to the delightful to the confusing. Aside from celebrities, many videos on Sora show convincing deepfakes of manipulated historical moments. <\/p>\n

For example, NBC News was able to generate realistic videos of President Dwight Eisenhower confessing to accepting millions of dollars in bribes, U.K. Prime Minister Margaret Thatcher arguing that the \u201cso-called D-Day landings\u201d were overblown, and President John F. Kennedy announcing that the moon landing was \u201cnot a triumph of science but a fabrication.\u201d<\/p>\n

The ability to generate such deepfakes of nonconsenting deceased individuals has already caused complaints from family members.<\/p>\n

In an Instagram story<\/a> posted Monday about Sora videos featuring Robin Williams, who died in 2014, Williams\u2019 daughter Zelda wrote: \u201cIf you\u2019ve got any decency, just stop doing this to him and to me, to everyone even, full stop. It\u2019s dumb, it\u2019s a waste of time and energy, and believe me, it\u2019s NOT what he\u2019d want.\u201d<\/p>\n

Bernice King, Martin Luther King Jr.\u2019s daughter, wrote on X<\/a>: \u201cI concur concerning my father. Please stop.\u201d King\u2019s famous \u201cI have a dream\u201d speech has been continuously manipulated and remixed on the app. <\/p>\n

George Carlin\u2019s daughter said in a BlueSky post<\/a> that his family was \u201cdoing our best to combat\u201d deepfakes of the late comedian.<\/p>\n

Sora-generated videos depicting \u201chorrific violence<\/a>\u201d involving renowned physicist Stephen Hawking have also surged in popularity this week, with many examples circulating on X<\/a>.<\/p>\n

A spokesperson for OpenAI told NBC News: \u201cWhile there are strong free speech interests in depicting historical figures, we believe that public figures and their families should ultimately have control over how their likeness is used. For public figures who are recently deceased, authorized representatives or owners of their estate can request that their likeness not be used in Sora cameos.\u201d<\/p>\n

In a blog post from last Friday<\/a>, OpenAI CEO Sam Altman wrote that the company would soon \u201cgive rightsholders more granular control over generation of characters,\u201d referring to wider types of content. \u201cWe are hearing from a lot of rightsholders who are very excited for this new kind of \u2018interactive fan fiction\u2019 and think this new kind of engagement will accrue a lot of value to them, but want the ability to specify how their characters can be used (including not at all).\u201d<\/p>\n

OpenAI\u2019s quickly evolving policies<\/a> for Sora have led some commentators to argue the company\u2019s move fast and break things<\/a> approach was purposeful, showing users and intellectual-property holders<\/a> the app\u2019s power and reach.<\/p>\n

Liam Mayes, a lecturer at Rice University\u2019s program in media studies, thinks increasingly realistic deepfakes could have two key societal effects. First, he said, \u201cwe\u2019ll find trusting people falling victim to all kinds of scams, big, powerful companies exerting coercive pressures and nefarious actors undermining democratic processes,\u201d Mayes said.<\/p>\n

At the same time, being unable to discern deepfakes from real video might reduce trust in genuine media. \u201cWe might see trust in all sorts of media establishments and institutions erode,\u201d Mayes said.<\/p>\n

As founder and chairman of CMG Worldwide, Mark Roesler has managed the intellectual property and licensing rights for over 3,000 deceased entertainment, sports, historical and music personalities like James Dean, Neil Armstrong and Albert Einstein. Roesler said that Sora is just the latest technology to raise concerns about protecting figures\u2019 legacies.<\/p>\n

\u201cThere is and will be abuse as there has always been with celebrities and their valuable intellectual property,\u201d he wrote in an email. \u201cWhen we began representing deceased personalities in 1981, the internet was not even in existence.\u201d<\/p>\n

\u201cNew technology and innovation help keep the legacies of many historical, iconic personalities alive, who shaped and influenced our history,\u201d Roesler added, saying that CMG will continue to represent its clients\u2019 interests within AI applications like Sora.<\/p>\n

To differentiate between a real and Sora-generated video, OpenAI implemented several tools to help users and digital platforms identify Sora-created content.<\/p>\n

Each video includes invisible signals, a visible watermark and metadata \u2014 behind-the-scenes technical information that describes the content as AI-generated.<\/p>\n

Yet several of these layers are easily removable, said Sid Srinivasan, a computer scientist at Harvard University. \u201cVisible watermarks and metadata will deter casual misuse through some friction, but they are easy enough to remove and won\u2019t stop more determined actors.\u201d<\/p>\n

Srinivasan said an invisible watermark and an associated detection tool would likely be the most reliable approach. \u201cUltimately, video-hosting platforms will likely need access to detection tools like this, and there\u2019s no clear timeline for wider access to such internal tools.\u201d<\/p>\n

Wenting Zheng, an assistant professor of computer science at Carnegie Mellon University, echoed that view, saying: \u201cTo automatically detect AI-generated materials on social media posts, it would be beneficial for OpenAI to share their tool for tracing images, audio and videos with the platforms to assist people in identifying AI-generated content.\u201d<\/p>\n

When asked for specifics about whether OpenAI had shared these detection tools with other platforms like Meta or X, a spokesperson from OpenAI referred NBC News to a general technical report<\/a>. The report does not provide such detailed information.<\/p>\n

To better identify genuine footage, some companies are resorting to AI to detect AI outputs, according to Ben Colman, CEO and co-founder of Reality Defender, a deepfake-detecting startup.<\/p>\n

\u201cHuman beings \u2014 even those trained on the problem, as some of our competitors are \u2014 are faulty and wrong, missing the unseeable or unhearable,\u201d Colman said.<\/p>\n

At Reality Defender, \u201cAI is used to detect AI,\u201d Colman told NBC News. AI-generated \u201cvideos may get more realistic to you and I, but AI can see and hear things that we cannot.\u201d<\/p>\n

Similarly, McAfee\u2019s Scam Detector software \u201clistens to a video\u2019s audio for AI fingerprints and analyzes it to determine whether the content is authentic or AI-generated,\u201d according to Steve Grobman, chief technology officer at McAfee.<\/p>\n

However, Grobman added, \u201cnew tools are making fake video and audio look more real all the time, and 1 in 5 people told us they or someone they know has already fallen victim to a deepfake scam.\u201d<\/p>\n

The quality of deepfakes also differs among languages, as current AI tools in commonly used languages like English, Spanish or Mandarin are vastly more capable<\/a> than tools in less commonly used languages.<\/p>\n

\u201cWe are regularly evolving the technology as new AI tools come out, and expanding beyond English so more languages and contexts are covered,\u201d Grobman said.<\/p>\n

Concerns about deepfakes have made headlines before. Less than a year ago, many observers predicted <\/a>that the 2024 elections would be overrun with deepfakes. This largely turned out not to be true<\/a>.<\/p>\n

Until this year, however, AI-generated media, like images, audio and video, has largely been distinguishable from real content. Many<\/a> commentators<\/a> have found models released in 2025 to be particularly lifelike, threatening the public\u2019s ability to discern real, human-created information from AI-generated content.<\/p>\n

Google\u2019s Veo 3 video-generation model, released in May, was called \u201cterrifyingly accurate<\/a>\u201d and \u201cdangerously lifelike<\/a>\u201d at the time, inspiring one reviewer to ask, \u201cAre we doomed?<\/a>\u201d<\/p>\n