生成せいAIエーアイ「Grok」による無断むだん性的せいてき画像がぞう生成せいせい問題もんだい、国際的こくさいてきな規制きせいと法的ほうてき責任せきにんをめぐる動向どうこう

Although the generative AI Grok does not permit the creation of pornographic images, problems caused by this AI, which is integrated into Elon Musks SNS X formerly Twitter, have become a serious social issue both domestically and internationally.

In particular, the spread of digitally altered images of real women with their clothes removed or in provocative poses on the internet is causing significant controversy from both ethical and legal perspectives.

Furthermore, images that appear to include individuals who look like minors have also been found, raising concerns that this may violate laws prohibiting child pornography.

The unlimited generation of images through the combination of AI and social media poses risks such as the violation of the rights of socially vulnerable individuals and potential breaches of domestic and international laws, yet in reality, sufficient countermeasures have not yet been implemented.

Mr. Musk and the xAI development company explain that they are taking measures such as removal of content, permanent suspension of accounts, and cooperation with authorities when necessary against illegal content, especially child sexual abuse material CSAM, but the generation of pornographic images by Grok has still not been stopped.

Mr. Musk has so far expressed opposition to regulations and censorship known as woke, and is said to be not proactive about strengthening oversight of Grok within the company.

On the other hand, this companys security team is smaller compared to other companies, and the fact that many employees resigned just before the problem became apparent is also cited as a reason for the delayed response.

Unlike other mainstream AI models, Grok sometimes allows, and even encourages, pornographic content and companion avatars.

Furthermore, since it is integrated with X, one of the worlds largest social networks, users can not only use Grok in private conversations, but also tag Grok in public posts to get answers, making the spread of problems even more serious.

Since the end of December 2023, there has been a rapid increase in digital undressing images created by Grok. Initially, these images mainly featured celebrities in bikinis, but recently, the number of prompts targeting ordinary women without their consent has also been rising.

According to research by European AI detection organizations and non-profit groups, words such as she, undress, and bikini frequently appear, with more than half of the generated images featuring underwear or bikini-clad figures, and over 80% depicting women.

Furthermore, it is said that 2% of the images contain subjects under the age of 18.

The terms of service of X and xAI clearly prohibit depicting individuals likenesses in a sexualized or pornographic manner involving minors, but due to operational limitations, illegal or inappropriate content continues to circulate.

In response to this situation, users have criticized that providing features without appropriate measures for children is irresponsible, and xAI has stated that they are considering strengthening their countermeasures.

Furthermore, Mr. Musk himself has issued a warning regarding illegal activities, stating, If you use Grok to create illegal content, the poster will receive the same level of punishment.

The issue of AI generating pornographic images without the consent of minors is not limited to Grok, but is also spreading to other platforms such as TikTok and OpenAIs Sora.

Researchers point out that, technically, it is possible to build a system in which AI detects children in images and prompts careful responses. However, challenges such as cost, response delays, and misrecognition still remain.

Currently, authorities in Europe, India, Malaysia, and other regions have begun investigating the issue of Groks image generation, and the United States is also considering pursuing legal responsibility.

According to experts at Stanford Universitys HAI Institute, companies operating social networks are partially exempt under Section 230 of the Communications Decency Act, but this does not apply to federal crimes including CSAM, and the possibility of being sued in civil court cannot be ruled out.

Due to the series of issues regarding Grok this time, there is a possibility that xAI will be perceived as more of a bare deepfake site rather than a major AI company.

As AI technology and social networks become increasingly interconnected, it goes without saying that it is an urgent issue to establish legal frameworks and ethics, and to take effective measures to fulfill social responsibility.