EXCLUSIVE: Celebrity 深い 偽の porn （テニスなどの）ダブルス in a year thanks to sophisticated AI as 星／主役にするs 含むing Taylor Swift, Natalie Portman and Emma Watson are 的d

• e-mail

Activity on 会議s 献身的な to celebrity 深い 偽の porn has almost 二塁打d in a year as sophisticated 人工的な 知能 (AI) becomes 広範囲にわたって 利用できる to the public, DailyMail.com can 明らかにする／漏らす.

Shockingly, the spike in perverted activity was （悪事，秘密などを）発見するd on websites easily accessible through Google and other internet browsers, meaning knowledge of how to use the dark web is not needed to 満足させる depraved fantasies.

A raft of high-profile 女性（の） 星／主役にするs, 含むing Taylor Swift, Natalie Portman and Emma Watson, have already had their images manipulated by 科学（工学）技術 to make them appear in erotic or pornographic content.

Now, a 報告（する）／憶測 by online safety 会社／堅い ActiveFence, 株d 排他的に with DailyMail.com, 解除するs the lid on the 新たな展開d world of 深い 偽の porn 会議s where perverts 株 online 道具s t hat can turn almost any celebrity picture into pornography.

場所／位置s are littered with 誇るs from 使用者s that their 科学（工学）技術 'helps to undress anybody', 同様に as guides on how to create 性の 構成要素, 含むing advice on what type of pictures to use.

A shocking 報告（する）／憶測 by online safety 会社／堅い ActiveFence, 株d 排他的に with DailyMail.com, has 解除するd the lid on the depraved world of 深い 偽の pornography 会議s

会議s discussing and 株ing 'celebrity 偽の photos' have seen activity rise by 87 パーセント since last year, with singer Taylor Swift a たびたび（訪れる） 的

深い 偽の celebrity porn is nothing new. In 2018, an image of actress Natalie Portman was computer-生成するd from hundreds of stills and featured in an explicit ビデオ

Harry Potter 星／主役にする Emma Watson has also been a popular 的 of 偽の porn perverts, but 専門家s say it is the 緩和する and 速度(を上げる） with which 科学（工学）技術 can now create manipulated explicit content that is of growing 関心?

The にわか景気 has created an entire 商業の 産業 built around 深い 偽の pornography, 含むing websites that have hundreds of thousands of 支払う／賃金ing members.

One of the most popular, MrDeepFakes, has around 17 million 訪問者s a month, によれば web analytics 会社／堅い SimilarWeb.

ActiveFence said the number of 会議s on the open web discussing or 株ing celebrity 深い 偽の porn between February and August this year was up by 87 per cent compared to the same period last year.

But 研究員s said 'no one is 安全な' from having their images 侵害する／違反するd, with the 百分率 増加する standing at a staggering 400 per cent for 私的な individuals.

It has 誘発するd 関心s that thousands could 落ちる 犠牲者 to AI-生成するd '復讐 porn'.

'The AI にわか景気'

深い 偽の porn is usually made by taking the 直面する of a person or celebrity and superimposing it の上に someone else's 団体／死体 成し遂げるing a 性の 行為／法令／行動する.

以前, a 使用者 would need technical 専門的知識 to create this content, with images taken from さまざまな different angles and photoshopping 技術s.

Now, all someone needs is a 非，不，無-nude image of their 犠牲者, often pilfered from social マスコミ or dating profiles, to 料金d into a chatbot.

This, in part, is thanks to the advent of generative AI - a form of AI that can 現実に make things, such as words, sounds and images.

深い 偽の 会議s are littered with guides on how to create 性の content, 含むing advice on what type of pictures to use, as shown above

新たな展開d 使用者s 誇る that their 人工的な 知能 models can 'undress anybody'

The problem has also been 悪化させるd by the 解放（する） of the codes used to make these chatbots by tech 巨大（な）s.

Chatbots made by 会社／堅いs such as OpenAI, Microsoft and Google come with rigorous 保護（する）／緊急輸入制限s, designed to 妨げる them from 存在 used to produce malicious content.

But in February, Meta - the tech 巨大（な） which owns Facebook, Instagram and WhatsApp - decided to make their code public, 許すing amateur tech 長，率いるs to 引き裂く out these filters.

Smaller tech companies followed 控訴.

It is no coincidence that ActiveFence （悪事，秘密などを）発見するd a spike in 深い 偽の porn from February this year, a moment it 述べるs as 'The AI にわか景気'.

'No one is 安全な'

REVEALED: AI chatbots 'uncensored' for the good of humanity are 存在 used by pedophiles and perverts to 生成する child pornography and graphic 性の 乱用 fantasies?

Predators are 偉業／利用するing 前進するs in the 科学（工学）技術 to create horrific, 高度に-現実主義の child 乱用 構成要素...

宣伝

伝統的に, 女性（の） celebrities have been the 的 of 深い 偽の pornography.

As long ago as 2018, an image of actress Natalie Portman was computer-生成するd from hundreds of stills and featured in an explicit ビデオ, as was Harry Potter 星／主役にする Emma Watson.

深い 偽の ビデオs featuring singer Taylor Swift have been 見解（をとる）d hundreds of thousands of times.

But ActiveFence 研究員 Amir Oneli said the 速度(を上げる） and 容積／容量 at which AI can now create 深い 偽のs means the general public are ますます 落ちるing 犠牲者.

He said that in the past, when AI-image 世代 was slow and difficult, 使用者s 焦点（を合わせる）d on celebrity content that would go viral.

'Today, we are seeing that it is 影響する／感情ing 私的な individuals because it's so 即座の, so 急速な／放蕩な,' he 追加するd.

'The most 悲劇の thing is that no one is 安全な.'

The ActiveFence 報告（する）／憶測 also 述べるs a 'vibrant scene of guides' 利用できる on the open web, advising 使用者s on which types of images to use and how to 修正する them.

One recommends pictures that show the 犠牲者 in a simple position, with their 団体／死体 明確に 陳列する，発揮するd, without baggy 着せる／賦与するing, and with a 'good contrast between 肌 and 着せる／賦与するing colors'.

深い 偽の chatbots 簡単に 教える 使用者s to 'select the photo you want to undress', while others （人命などを）奪う，主張する 'its 前進するd image editing 科学（工学）技術 can easily 除去する 着せる／賦与するing from any picture, leaving behind only the 明らかにする 必須のs'.

以前, creating 深い 偽の porn 要求するd technical 技術, but it is now as simple as feeding an image of the 的d individual into a chatbot, which will turn it into explicit content

地位，任命するs on 深い 偽の 会議s encourage 使用者s to create their own 構成要素 by explaining how simple it is. One says 'finally there is a 簡単に way to 除去する 着せる／賦与するs from any photo'

There is now an entire 産業 built around 深い 偽の porn, with one website 非難する up to $560 for AI-生成するd explicit content and 申し込む／申し出ing 使用者s the chance to 'skip the 列' for downloads if they cough up even more?

Cashing in on depravity

Bots and websites that 許す 使用者s to create nude images at 速度(を上げる） まず第一に／本来 operate on a subscription model, 許すing 使用者s to either 支払う／賃金 a 月毎の or a per use 料金 for each image.

Some 告発（する），告訴（する）／料金 as little as $6 to download 偽の porn pictures, with prices going up to $560 for 2,000 images, ActiveFence 設立する.

The 需要・要求する is such that some websites 申し込む／申し出 '急速な／放蕩な pass' tickets 許すing u sers to 支払う／賃金 $400 to 'skip the 列' to 接近 images, with the ticket 満了する／死ぬing after nine hours.

MrDeepFakes is one of the most popular 深い 偽の porn websites, and comes out 最高の，を越す of a Google search for the 称する，呼ぶ／期間／用語.

The website hosts short teaser ビデオs, which entice 使用者s to buy longer 見解／翻訳／版s from another website: Fan-Topia, によれば a 最近の 調査 by NBC.

Fan-Topia 述べるs itself as the highest 支払う／賃金ing adult content creator 壇・綱領・公約.

Noelle ツバメ, a 合法的な 専門家 on 科学（工学）技術-容易にするd 性の 乱用, told NBC MrDeepFakes was 'not a porn 場所／位置' but a 'predatory wesbite that doesn't rely on the 同意 of the people on the actual website'.

'The fact that it's even 許すd to operate and is known is a 完全にする 起訴，告発 of every regulator in the space, of all 法律 施行, of the entire system,' she told NBC.

非，不，無-consensual 株ing of sexually explicit imagery is 違法な in most 明言する／公表するs, but this does not 適用する to deepfake 構成要素 in all but four 明言する／公表するs: California, Georgia, New York and Virginia.

The 妨げるing Deepfakes of Intimate Images 行為／法令／行動する was introduced to 議会 in May in a 企て，努力，提案 to make 株ing 非，不，無-consensual AI-生成するd porn 違法な in the US, but it has yet to be passed.