Breaking AI & Tech News Analyzed

YouTube is significantly expanding its AI likeness detection technology, moving beyond individual creators and politicians to target the broader entertainment industry. Announced on Tuesday, this expansion aims to protect celebrities, talent agencies, and management companies from the unauthorized use of their digital identities in AI-generated content, such as deepfakes and scam advertisements.Key DevelopmentsPilot Phase: The technology was first tested with a subset of creators last year before expanding to politicians and government officials in the spring.Industry Rollout: The tool is now available to talent agencies, management companies, and the celebrities they represent, including major industry players like CAA, UTA, WME, and Untitled Management.Operational Mechanics: Unlike previous iterations, users do not need their own YouTube channels to utilize the tool. The system scans for visual matches of an enrolled participant’s face and offers options to request removal for privacy violations or submit a copyright claim.Future Scope: YouTube announced that audio support will be added to the technology in the future, broadening its capability to detect synthetic voice clones.Data & Market ImpactWhile YouTube has not disclosed the exact number of removals managed by the tool to date, the company noted in March that the volume of AI deepfake removals was still “very small.” This indicates that while the technology is live, the prevalence of high-quality, malicious deepfakes targeting celebrities is currently in its early stages. However, the strategic partnership with top-tier agencies signals a massive shift in market dynamics, treating digital likeness as a high-value asset comparable to intellectual property.Why This MattersThis expansion is critical for the entertainment industry because it addresses a vulnerability that traditional copyright laws struggle to cover. Celebrities frequently find their likenesses used in scam advertisements or non-consensual content, causing severe reputational damage and financial loss. By providing a technical solution that operates similarly to Content ID, YouTube is effectively creating a new standard for digital rights management in the age of generative AI. This move protects not just individual stars but the entire ecosystem of talent management.Expert InsightThe integration of major agencies like CAA and UTA into the pilot program validates the necessity of automated detection tools. Unlike copyright, which protects expression, likeness protection is about identity. The fact that top-tier agencies are adopting this tech suggests a proactive approach to risk management. It also highlights a strategic pivot for YouTube: moving from a platform that hosts content to a platform that actively polices the integrity of the digital identities represented on it. This partnership likely provides YouTube with valuable feedback on how to refine the algorithm to distinguish between malicious deepfakes and permissible parody or satire.What Happens NextWe can expect the technology to evolve rapidly, particularly with the upcoming addition of audio detection. As generative AI becomes more accessible, the volume of unauthorized content will likely increase, prompting YouTube to refine its detection accuracy. Furthermore, the success of this tool may accelerate the passage of the NO FAKES Act in Washington, D.C., as industry stakeholders gain a technical foothold in the fight against synthetic media. The battle between AI creators and detection systems will likely intensify, making this a defining feature of the platform's future policy landscape.

As the London Marathon approaches on 26 April, runners are being lured by fake offers to buy a race place for £79 via bank transfer – a scam that exploits the high demand for the coveted ballot entry.Key DevelopmentsScammers post in running‑app groups claiming injury and offering to "sell" a marathon slot for £79 via bank transfer.The official organisers state that marathon entries are strictly non‑transferable under any circumstances.Victims are asked to provide full name, email and payment details, mirroring the legitimate entry fee of £79.99.Red flags include poor grammar, bank‑transfer requests, and the promise of a quick bib transfer on the marathon website.Strava has warned that such activity breaches its policies and will result in account suspension.Data & Market ImpactEntry fee for a legitimate London Marathon spot: £79.99.Scam fee demanded: £79, a near‑identical amount designed to lower suspicion.Potential loss per victim: up to £79, plus possible exposure of personal banking details.With over 40,000 runners applying annually, even a 0.1% fraud conversion would affect dozens of participants and erode trust in official channels.Why This MattersRunning enthusiasts and charity fundraisers rely on the integrity of the ballot system. Fraudulent offers not only risk financial loss for individuals but also threaten the reputation of the event, which raises millions for charity. The use of bank transfers bypasses consumer protections such as credit‑card chargeback rights, leaving victims with limited recourse.Expert InsightEvent‑ticket scams spike when demand peaks and official supply is limited. The London Marathon model—ballot entry, non‑transferable bibs, and a modest fee—creates a perfect lure for fraudsters who mimic official language. The reliance on third‑party apps like Strava amplifies the problem, as community groups lack verification mechanisms. Regulators and organisers must combine clear communication with technical safeguards (e.g., verified seller badges) to curb the abuse.What Happens NextOrganisers will likely intensify public warnings through the marathon website and partner apps.Strava may introduce stricter monitoring of marketplace‑style posts and expand its reporting tools.Potential legislative pressure could lead to tighter rules on the sale of non‑transferable event tickets in the UK.Runners are advised to stick to official ballot entries or charity slots and to avoid any payment method that lacks consumer protection.

For the first time in Hollywood history, an authorized generative‑AI version of a major star headlines a feature film. The western As Deep As the Grave showcases a digitally recreated Val Kilmer, whose voice was synthesized by UK‑based firm Sonantic using archival recordings. The project, delayed by Kilmer’s death in April 2025, received full cooperation from his estate and his daughter Mercedes, who helped craft the visual deep‑fake of the actor. Mercedes Kilmer confirmed that her father embraced emerging technologies as a storytelling tool, a sentiment the filmmakers say they honored throughout production. At Wednesday’s CinemaCon trade show in Las Vegas, the trailer revealed that Kilmer’s character, Father Fintan—a Catholic priest and Native American spiritualist—appears for roughly one hour of screen time. The footage shows the priest at different ages, including a spectral elder who advises a child, “Don’t fear the dead and don’t fear me.” Writer‑director Coerte Voorhees explained that the narrative was built around Kilmer’s heritage and his love of the Southwest. He added that the production adhered to SAG‑AFTRA guidelines and that the estate was financially compensated for the use of archival material. The film also stars Tom Felton, Abigail Breslin and Abigail Lawrie. In a March interview with Variety, Voorhees noted that Kilmer’s family repeatedly emphasized the project’s importance to the late actor. The Kilmer case follows a growing trend of AI‑generated performances. In 2022, Bruce Willis consented to a digital twin after a dementia diagnosis, while actors such as Matthew McConaughey and Michael Caine have licensed their voices to AI firms for approved uses. Estates of legends like Laurence Olivier, Judy Garland and James Dean have similarly partnered with the marketplace ElevenLabs. Beyond film, celebrities are exploring AI for digital meet‑and‑greets; Paris Hilton and Kendall Jenner have reportedly signed deals with Meta for AI‑powered appearances on Instagram, and users can even query an AI version of Deepak Chopra for advice on a range of topics. Nevertheless, some stars remain cautious. Morgan Freeman, Tom Hanks and Scarlett Johansson have publicly expressed skepticism about deep‑fake replicas of their likenesses.

A former Alabama defensive lineman, Luther Davis, who helped the Crimson Tide win the 2010 national championship, is preparing to plead guilty to a multi‑million‑dollar loan fraud that hinged on impersonating NFL athletes. According to a criminal information filing by the U.S. Attorney for the Northern District of Georgia, Davis and his associate CJ Evins obtained at least thirteen fraudulent loans totaling $19,845,000. The defendants chose to waive a grand‑jury indictment and will enter guilty pleas at a hearing scheduled for 27 April. The scheme targeted lenders that specialize in financing athletes, notably Aliya Sports and All Pro Capital Funding, with loan brokerage services provided by Sure Sports. Three of the loans are detailed in the filing: $4.025 million was secured for a fictitious company linked to Cleveland Browns tight end David Njoku. $4.35 million was obtained for a sham entity tied to Green Bay Packers safety Xavier McKinney. $3.3 million was borrowed for a fabricated venture associated with Atlanta Falcons quarterback Michael Penix Jr. Investigators say the duo created shell companies with names resembling the players’ initials, opened bank accounts, and fabricated email addresses and driver’s licenses. Davis then attended virtual loan closings in disguise—often wearing wigs, makeup, or a durag—to pose as the athletes and convince notaries to certify the fraudulent documents. One closing on 22 January 2024 for the Njoku loan involved Davis presenting a counterfeit Georgia driver’s license that displayed the player’s photo alongside a number belonging to an unrelated Savannah resident. Similar deceptions occurred for the McKinney and Penix loans, with forged Florida and Georgia licenses respectively. The fraud mirrors a separate case in which First Farmers Bank & Trust sued an insurer after a $5.265 million loan, also brokered by Sure Sports, was discovered to have been signed with a fake Njoku identity. While it is unclear whether that loan is part of the thirteen identified in Georgia, the modus operandi aligns closely. Both Davis and Evins face charges of aggravated identity theft and conspiracy to commit wire fraud, offenses that carry potential sentences of up to 20 years in prison. Their attorneys declined to comment on the pending pleas. Beyond the courtroom, the case underscores vulnerabilities in niche financing markets that cater to professional athletes, highlighting how forged identities and shell corporations can be leveraged to extract substantial capital from lenders.

Swiss ice‑hockey head coach Patrick Fischer has publicly acknowledged that he travelled to Beijing for the 2022 Winter Olympics using a fabricated COVID‑19 vaccination certificate. In a statement, Fischer described the act as a “serious mistake” and expressed regret for disappointing fans and officials. Fischer explained that he was caught in a personal dilemma, refusing vaccination yet unwilling to let his team miss the Games. "I was in an extraordinary personal crisis because I didn’t want to be vaccinated," he said, adding that he did not want to let his squad down. Swiss public broadcaster SRF presented evidence that Fischer had been fined nearly 39,000 Swiss francs (about $50,000) in 2023 for document forgery after purchasing the counterfeit certificate on social media. The coach disclosed his admission shortly after SRF’s confrontation. Despite the controversy, Fischer remains at the helm of the Swiss national team until after the World Championship scheduled next month, after which he is set to step down. The Swiss Ice Hockey Federation has stated that it considers the matter closed. Fischer, who has led the team since 2015, is among Switzerland’s most successful hockey coaches, guiding the squad to three Olympic appearances and securing three silver medals at the World Championships. Under his leadership, Switzerland reached the quarterfinals at the 2022 Olympics, a tournament that required strict COVID‑19 testing and saw the NHL abstain due to pandemic concerns. China imposed some of the world’s toughest COVID‑19 protocols for the Beijing Games, mandating vaccination or a three‑week quarantine for all athletes—a rule that Swiss snowboarder Patrizia Kummer chose to follow. The International Olympic Committee has not yet commented on the incident.

A Miami‑based U.S. District Judge, Darrin Gayles, dismissed former President Donald Trump’s $10 billion defamation lawsuit against the Wall Street Journal and its proprietor Rupert Murdoch. The case centered on a July 2025 article that linked Trump to a birthday greeting allegedly sent to convicted sex offender Jeffrey Epstein.Judge Gayles concluded that Trump, as a public figure, did not satisfy the stringent “actual malice” threshold required in defamation actions. To prevail, a plaintiff must prove that the media outlet knowingly published false information or acted with reckless disregard for the truth.In his written opinion, Gayles noted that WSJ reporters had reached out to Trump for comment before publishing the story and included his denial, thereby giving readers a balanced view. He wrote, "This complaint comes nowhere close to the actual‑malice standard—quite the opposite."The judge granted Trump permission to file an amended complaint, setting a deadline of April 27 for any revisions.Trump’s original filing labeled the alleged birthday note to Epstein as a “fake” and sought damages for perceived harm to his reputation. The newspaper’s parent company, News Corp’s Dow Jones & Company, defended the article’s accuracy, emphasizing its adherence to journalistic standards.The dismissal adds to a series of legal setbacks for the former president as he attempts to curb reporting on his connections to Epstein. Trump announced on his Truth Social platform that he intends to re‑file the suit within the court‑ordered timeframe.A Dow Jones spokesperson welcomed the decision, stating, "We are pleased with the judge’s decision to dismiss this complaint and stand behind the reliability, rigor, and accuracy of The Wall Street Journal’s reporting."

A federal judge in Miami has dismissed former President Donald Trump’s defamation lawsuit against the Wall Street Journal and News Corp, granting the former president a two‑week deadline to refile the case. Trump’s suit, filed last summer, alleged that a lewd drawing featured in a July 2025 article—purportedly a “bawdy” birthday letter to the late financier Jeffrey Epstein—was fabricated, and that the newspaper published it with actual malice. The complaint also named media mogul Rupert Murdoch, whose News Corp owns the Journal, as a defendant. Judge Darrin P. Gayles ruled that the complaint “fails to adequately allege actual malice,” the legal standard required for defamation actions by public figures. He noted that the Journal had conducted a “significant” inquiry into the authenticity of the drawing and that Trump’s assertion of falsity alone does not prove the newspaper acted with “serious doubts” about the story’s truth. In his opinion, the judge wrote: “Because President Trump has not plausibly alleged that defendants published the article with actual malice, both counts must be dismissed.” He also observed that Trump’s team had not presented evidence of special damages. Under the order, Trump may refile the lawsuit by April 27 with additional proof that the Journal knowingly published false material. A spokesperson for Trump’s legal team confirmed they will pursue a revised filing, emphasizing the administration’s intent to “hold accountable those who traffic in fake news.” The Wall Street Journal and its parent company, Dow Jones, welcomed the decision. A Dow Jones representative said, “We stand behind the reliability, rigor and accuracy of The Wall Street Journal’s reporting.” The dismissal underscores the stringent “actual malice” requirement for defamation suits involving public officials, a threshold that continues to shape media‑law battles in the United States. It also leaves Trump with ongoing litigation against the BBC and other media outlets over separate First Amendment disputes.

Booking.com, a leading accommodation reservation website, has suffered a significant data breach that has exposed customer information to unauthorized parties. The company, which lists over 30 million accommodation venues worldwide, detected suspicious activity involving unauthorized access to some guests' booking information.Upon discovering the breach, Booking.com took immediate action to contain the issue and updated the PIN numbers for affected reservations. The company has also informed affected customers about the breach. According to Booking.com, financial information was not accessed during the breach.The breach is the latest in a series of cybercrime attempts on Booking.com, which has recently struggled with a rising number of online scams on its platform. In 2018, the company reported a breach that exposed the booking data of over 4,000 people. Booking.com was fined €475,000 for reporting the breach 22 days late to the Dutch privacy regulator.The company, owned by Booking Holdings, a $137 billion US company, employs over 24,000 people worldwide. The breach highlights the growing concern of fake listings on booking websites and the need for increased cybersecurity measures in the industry.

Renowned jazz composer and pianist Jason Moran recently discovered a fake album on Spotify bearing his name. The album, titled For You, had a moody Japanese anime-style cover and featured indie pop music, which was a far cry from Moran's actual work.Moran is not alone in this experience. At least a dozen famous musicians, including Benny Green, Antonio Hart, and Dee Dee Bridgewater, have been targeted by AI bots impersonating them on music streaming platforms. The issue has led to frustration and surreal experiences for the musicians, who are now having to deal with the deluge of AI-generated music.Spotify has acknowledged the problem and has taken steps to address it, including removing over 75 million "spammy tracks" from its platform in the past year. The company is also working on a new tool to give artists more control over what shows up under their name.However, for musicians like Moran, these fixes aren't enough. He's concerned about the additional work for artists who don't put their music on Spotify, and for musicians who are no longer alive. Morgan Hayduk, a co-CEO of Beatdapp, estimates that 5% to 10% of all streams across the industry are fraudulent, which breaks down to a value of $1 billion to $2 billion per year.The issue highlights the challenges of regulating AI-generated content and the need for more effective solutions to prevent music impersonation on streaming platforms.