During a recent signing ceremony, California Governor Gavin Newsom announced a series of new laws aimed at the artificial intelligence (AI) industry, which are relatively strict on a national level. The new regulations comprise five items, three of which specifically address AI deepfake technology that could impact elections, while the other two concern the cloning of Hollywood actors' voices and images. Image source note: Image generated by AI, image authorized by service provider Midjourney. The new law AB2655 requires large online platforms...

Recently, several major American artificial intelligence companies made a solemn pledge to the White House, stating that they will prevent their AI products from being used to generate non-consensual deepfake pornography and child sexual abuse materials. Six companies, including Adobe, Anthropic, Cohere, Microsoft, OpenAI, and the open-source web data repository Common Crawl, signed this non-binding commitment to ensure that their products are not misused to create harmful images of gender violence.

In today's world, deepfakes and synthetic media have become increasingly common due to the advancements in generative artificial intelligence (gen AI). To tackle this challenge, many organizations are working to improve digital watermarking technologies for tracking the origins of content. While no organization has completely succeeded yet, the Coalition for Content Provenance and Authenticity (C2PA) is expanding its membership. C2PA has announced that Amazon has become a new member of the organization's advisory board. C2PA is an open technical standard organization under the Linux Foundation.

With the continuous advancement of artificial intelligence technology, the difficulty of forging images and videos has decreased significantly, making the deepfake phenomenon increasingly severe. Identifying these false contents has become an urgent issue to address. Recently, a research team from Binghamton University conducted an in-depth discussion on this topic, using frequency domain analysis techniques to reveal the characteristics of AI-generated images, thus helping people identify false information. Image source note: The image is AI-generated, with permission from image licensing provider Midjourney. This research was conducted by the Department of Electrical and Computer Engineering.

Singapore has recently introduced a new bill aimed at banning the use of deepfake technology during elections to ensure the authenticity of information and the fairness of the electoral process. With the advancement of generative AI technology, the risk of deepfake content has increased. The bill seeks to regulate the production of false content created by both AI and non-AI technologies, including videos, audio, and images. The core provisions prohibit the publication of online ads that falsely depict candidates' words and actions during election periods, with violators facing fines or imprisonment. The bill also allows for issuing correction orders to individuals and social media platforms that publish false content.

Microsoft collaborates with the anti-revenge pornography organization StopNCII to provide victims with tools to erase deepfake pornographic images from search engines. By creating a digital fingerprint (hash value) of the images on the device, it helps to completely remove related synthetic nude images from platforms like Bing. This initiative aims to address the issue of deepfake pornography caused by the misuse of generative AI technology, tackling the challenges that cannot be effectively managed solely through user reports. However, Google has not yet partnered with StopNCII, and there is a lack of specific federal laws addressing such behavior in the United States.

As the 2024 U.S. presidential election approaches, generative artificial intelligence (AI) is increasingly becoming a new player on the electoral battlefield. AI technology can generate false images, videos, and audio, influencing voter decision-making, a trend that has manifested in multiple instances. For example, former President Trump shared AI-generated images attempting to mislead the public, while residents of New Hampshire encountered deepfake calls aimed at dissuading them from participating in the Democratic primaries. Experts warn that the impact of AI-generated misinformation could substantially alter election outcomes, and current detection technologies remain underdeveloped.

The issue of South Korean teenagers using deepfake technology to create pornographic content has garnered widespread attention, with a surge in online sexual crimes. Police are urging for enhanced education and guidance for youth to improve online safety and moral awareness. Instances of pornographic content targeting schools and teachers have emerged on Telegram, with multiple cases reported. In the first seven months of this year, the number of....

The issue of deepfake crimes in South Korea has drawn widespread attention from society, particularly concerning the increasing number of cases involving minors, with an estimated 220,000 individuals affected. The victim demographics are broad, including students, teachers, military personnel, and a significant proportion of minors. These crimes are rapidly disseminated through social media, with some victims' information being used for validation in illicit groups. Cases of deepfake targeting women have garnered particular attention, with known affected schools exceeding 100, including both domestic and overseas Korean international schools. In response, the South Korean president has called for strengthened investigations and the eradication of crimes utilizing digital technology.

The Federal Communications Commission (FCC) has fined Lingo Telecom $1 million for distributing a fabricated voice call of President Joe Biden, aimed at influencing voters in New Hampshire. The investigation revealed that the mastermind behind the call was political consultant Steve Kramer, who the FCC had previously proposed a $6 million fine against. To reach a settlement with Lingo, the FCC has required the company to strictly adhere to caller ID authentication rules going forward, including implementing 'Know Your Customer' (KYC) measures.

San Francisco's Deputy Mayor Yvonne Meré has initiated a lawsuit against 16 websites that exploit AI technology to create deepfake pornographic content without the consent of women and girls. The lawsuit addresses the harmful trend among teenagers, particularly regarding the use of "strip apps" to manipulate peers' images. These websites garnered 200 million visits in their first six months, offering free initial services followed by paid options....

Recently, Google has taken action to combat non-consensual deepfake content in response to public concern and legislative pressure. Following the Taylor Swift incident eight months ago, tech companies and lawmakers began seriously addressing this issue. Google announced measures to prevent pornographic deepfake content from appearing in search results, including streamlining the process for victims to remove non-consensual pornography, filtering pornographic search results, removing duplicate images, and optimizing search results to reduce the exposure of deepfake content. Henry Ajdah has expressed support for Google's initiatives, believing it will significantly reduce the exposure of non-consensual content.

Google has launched a series of online safety measures aimed at eliminating deepfake content from search results, particularly explicit images without consent. When users request removal of such content, the system not only processes the specific item but also filters out related results and even removes duplicates. To prevent such content from appearing prominently in search results, Google has adjusted its ranking algorithm to prioritize high-q....

Microsoft Vice Chairman Brad Smith urged the U.S. Congress to implement laws targeting AI-generated deepfake technology to protect citizens from fraud and manipulation. He pointed out that while the industry and nonprofit organizations have taken measures to address this issue, laws need to be updated to combat deepfake fraud. Smith emphasized the need for a comprehensive 'Deepfake Fraud Act' to provide a legal framework for law enforcement to prosecute cybercriminals who exploit the technology for fraud. He also called for updating laws concerning child sexual abuse and non-consensual intimate images to include those generated by AI.

The U.S. Senate unanimously passed the 'Disrupt Explicit Forged Images and Non-Consensual Edits Act' aimed at combating non-consensual deepfake pornography, providing legal avenues for victims to hold perpetrators accountable. The bill grants victims the right to sue individuals or organizations that create or distribute fake images, with damages potentially reaching up to $250,000, aiming to protect public figures and teenagers from such issues.

Telecom network fraud methods are constantly evolving, posing a threat to public property security. Major types include 'screen sharing' fraud, which involves inducing victims to enable screen sharing to obtain personal banking information for theft; 'AI deepfake' fraud, which utilizes technology to synthesize false audio and video, gaining trust before misleading victims into transferring money or providing sensitive information; false online investment fraud that uses high returns as bait to lure individuals into investing before siphoning off funds; and fake transaction fraud in online gaming products, which deceives players into making offline transactions through fictitious trading information, resulting in financial losses. Prevention measures are necessary.

As one of the fastest-growing forms of adversarial artificial intelligence, the losses related to Deepfakes are expected to skyrocket from $12.3 billion in 2023 to $40 billion by 2027, with a compound annual growth rate of an astonishing 32%. Deloitte predicts a surge in deepfakes in the coming years, with banks and financial services becoming primary targets. Deepfakes represent the cutting edge of adversarial AI attacks, with a growth rate of 3,000% just last year. It is projected that Deepfak

Resemble AI, a leading voice cloning company, recently announced the launch of its next-generation deepfake detection model, Detect-2B. This new model has demonstrated approximately 94% accuracy in detecting AI-generated audio, marking another significant advancement in deepfake detection technology.

Google DeepMind has conducted research on the most common malicious applications of AI for the first time, revealing that AI-generated 'deepfakes' of false politicians and celebrities are more prevalent than attempts to use AI-assisted cyber attacks. This study was a collaboration between Google's AI division DeepMind and its R&D unit Jigsaw, aimed at quantifying the risks of AI generation tools that have been pushed to market by the world's largest tech companies seeking massive profits.