Latest news with #Sensity
Khaleej Times
15-04-2025
- Khaleej Times
Japanese police arrest four for selling obscene AI images
Japanese police have arrested four people for selling obscene images created using generative AI in the first crackdown of its kind, a police spokesman and local media reports said on Tuesday. The four, who allegedly made posters and sold them online, "were arrested on Monday on suspicion of selling obscene images", a Tokyo police spokesman told AFP. They sold them on auction sites several times last October, criminal acts for which they face up to two years in prison, fines of up to 2.5 million yen ($17,500), or both, he said. Public broadcaster NHK and other media outlets said the suspects used free AI software to create images of naked adult women who do not exist in the real world. They used prompts including terms such as "legs open". The four, aged in their 20s to 50s, reportedly sold the posters for several thousand yen (around $20-50) each. Tuesday's reports said the arrests were Japan's first for allegedly selling AI-created obscene images, which police could not immediately confirm. Concern is growing worldwide over the use of AI for malicious purposes including through deepfakes, which turn genuine photos, video or audio of people into false likenesses. Around 96 percent of deepfake videos online are non-consensual pornography, and most of them depict women, according to a 2019 study by the Dutch AI company Sensity.
The Independent
15-04-2025
- The Independent
Four arrested for selling obscene AI images in first crackdown of its kind
Japanese police have arrested four people for selling AI -generated obscene images in the first of its kind clampdown on the sale of pornographic images in the country. The suspects, arrested in Tokyo, were accused of selling posters with obscene images on internet auction sites in October of last year, NHK reported. They had allegedly used AI software to make nude images of non-existent women. The posters were reportedly sold for thousands of yen. A 2019 study by Dutch AI company Sensity found that around 96 per cent of deepfake videos online were non-consensual pornography. According to a survey by Japanese newspaper Yomiuri Shimbun, the East Asian nation ranks third in traffic to websites that disseminate sexually explicit deepfake images, with over 18 million visits. The United States ranks second, while India ranks first in terms of users visiting these websites. The survey, conducted with digital analytics firm Similarweb, analysed data from 41 websites that enabled users to produce sexually explicit images between December 2023 and November 2024. Nearly 410,000 users in Japan visited these websites monthly, with 80 per cent accessing through their smartphones. Digital technology experts have demanded better regulation of such content. "It is believed that there has potentially been much more damage than we are aware of,' said professor Yuasa Harumichi of Meiji University. 'Regulations on fake porn should be imposed before the damage spreads, and discussions should start immediately,' he told NHK in December 2024.
Japan Times
15-04-2025
- Japan Times
Four arrested over obscene AI images in Japan first: reports
Police have arrested four people for selling obscene images created using generative AI in the first crackdown of its kind, local media reports said Tuesday. The four, aged in their 20s to 50s, allegedly made posters featuring indecent images of women and sold them on internet auction sites, public broadcaster NHK and other outlets said, citing police sources. Police could not immediately confirm the reports. NHK said the suspects had used free AI software to create images of naked adult women, who do not exist in the real world, using prompts including terms such as "legs open". They reportedly sold the posters for several thousand yen (several multiples of $7) each. Tuesday's reports said the arrests were Japan's first for allegedly selling AI-created obscene images. Concern is growing worldwide over the use of AI for malicious purposes including through deepfakes, which turn genuine photos, video or audio of people into false likenesses. Around 96% of deepfake videos online are non-consensual pornography, and most of them depict women, according to a 2019 study by the Dutch AI company Sensity.
Yahoo
19-02-2025
- Entertainment
- Yahoo
AI-generated footage of zombie-like creatures does not show lab-grown humans
'America is producing human beings in laboratories,' claims the caption on a Facebook Reel written in Amharic, one of Ethiopia's major languages. Shared more than 150 times, the post contains a 15-second clip that appears to show zombie-like creatures thrashing around in a glass tank. When the creatures try to escape, people who are filming them panic and flee. 'We have enough crazy people, why is America creating more?' reads one of the comments under the Facebook post. 'If America has a shortage of people, why don't they get some from us instead of creating new ones,' adds another user. The footage, however, does not depict real events. AFP Fact Check analysed the footage using an AI detection tool called Sensity. The results show there is a 96 percent probability that the footage was made with an online generative tool (archived here). In addition, the analysis confirmed that the section of the footage purporting to show zombies in containment is entirely synthetic. AFP Fact Check also used the video verification tool InVID-WeVerify to conduct reverse image searches on keyframes from the footage. The search results include the original footage published on TikTok by a user called ' on December 15, 2024 (archived here). It was shared more than 400,000 times and has more than two million views. The original footage is accompanied by a caption that reads: 'When a Zombie Exhibition Goes Horribly Wrong.' 'Step into the chilling public exhibition at Area 51, where rows of zombie-like creatures float eerily in containment tanks,' adds the caption. AI-generated content often contains visual anomalies, particularly when it comes to mimicking human features (archived here). Several distortions occur in the footage: a member of the public morphs in and out of being a zombie while one of the mutant creatures appears to fly down from the ceiling and then merges with the crowd. AFP Fact Check further reviewed the TikTok account where the footage was originally posted and confirmed that it features numerous AI-generated clips of aliens and zombies purportedly found in Area 51. 'Inspiring creator on TikTok, sharing creativity and entertaining videos,' reads the bio. The account has more than 640,000 followers. Similar videos published by the same account indicate in the caption they have been created using AI (archived here). Area 51, officially known as the Nevada Test and Training Range, is a top-secret US Air Force facility at Groom Lake in southern Nevada (archived here). According to the US Air Force, it provides 'multidimensional battle-space to conduct testing tactics development, and advanced training in support of U.S. national interests' (archived here). The facility is not open to the public and is patrolled by armed guards. It is also impossible to enter the airspace above the facility without permission from air traffic control (archived here). Area 51 is conspiracy fodder for those who believe it harbours the remains of extraterrestrials and unidentified flying objects (archived here). AFP Fact Check has contacted the US facility for a comment and this story will be updated as the response becomes available.
