AI Library
The Age of Autonomous Scientific Discovery
Kim Kyung-jin, Attorney at Law
AI Scientists and Self-Driving Labs
This book follows how AI scientists and self-driving labs are changing the way science generates and verifies claims. It covers literature-based discovery, natural-language protocols translated into robot commands, multi-agent research systems, closed-loop laboratories, materials search, the verification gap, chains of evidence, research harnesses, journal ethics, and legal responsibility.
AI Library
A New Era of Life Sciences Opened by Artificial Intelligence
Structural Proteomics, Genomic Foundation Models, Autonomous Laboratories, and Global Governance
Kim Kyung-jin, Attorney at Law
This book is a research volume compiled with artificial intelligence. A human selected the materials and structured the work, while AI models drafted the sentences and cross-checked the facts.
AI Library
The Double Structure of Digital Sovereignty
Europe’s Departure from Palantir and the Chains of American Big Tech
Kim Kyung-jin, Attorney at Law
This is a record of 2026, when European intelligence agencies and defense ministries began removing analytics tools from America’s Palantir. It covers the replacement decisions made by France’s General Directorate for Internal Security (DGSI), Germany’s Federal Office for the Protection of the Constitution (BfV), and the Netherlands Ministry of Defense; the incident in which US export controls severed an ally’s ac…
New English Edition
Artificial Intelligence in Horticulture
Kim Kyung-jin, Attorney at Law
Across five chapters and ten sections, this book examines computer vision for crop diagnosis, harvesting robots and autonomous field systems, smart greenhouses and digital twins, precision irrigation and supply-chain quality control, high-throughput phenotyping, and predictive breeding.
New English Edition
Artificial Intelligence in Food Crop Agriculture
Kim Kyung-jin, Attorney at Law
Across six chapters and eighteen sections, the book examines digital agricultural infrastructure, remote sensing, crop diagnosis, yield forecasting, precision irrigation, genomics, molecular breeding, agricultural robotics, climate-smart agriculture, and global food security.
New English Edition
The Future of Forestry and Agroforestry
Kim Kyung-jin, Attorney at Law
Driven by Artificial Intelligence and Digital Innovation
Across five chapters and fifteen sections, the book follows satellites, drones, LiDAR, digital twins, forest-specific language models, wildfire and pest forecasting, forestry robotics, agroforestry, timber traceability, and forest carbon markets.
New English Edition
Smart Livestock Farming: AI Enters the Barn
Kim Kyung-jin, Attorney at Law
Sensors listen, cameras watch, and artificial intelligence helps farmers decide.
Across five chapters and fifteen sections, the book follows precision livestock farming from animal health and reproduction to robotic milking, virtual fencing, digital twins, methane reduction, welfare, and data ownership.
Table of Contents
Han Dong-hoon, Busan Buk-gu Gap: A Record of the 100 Days Before and After the Election (Mar. 26-Jul. 3, 2026)
Kim Kyung-jin
Table of Contents and 13 sections
From March 26 to July 3, 2026, this record follows the spring after expulsion, the Busan Buk-gu Gap by-election, victory as an independent, and the first bill submitted in the National Assembly.
Table of Contents
Artificial Intelligence and Medicine
Kim Kyung-jin, Attorney at Law
AI in clinical care, hospitals, education, and research
AI in medical imaging, risk prediction, treatment planning, hospital operations, education, and research, with patient safety, privacy, and accountability.
[AI Library] Chapter 9. Deepfakes and Synthetic Media Crimes
Artificial Intelligence on Trial
Part 3. Stolen Faces and Voices in the Digital World
Chapter 9. Deepfakes and Synthetic Media Crimes
Attorney Kyungjin Kim
A. The Fake Videos That Swallowed Taylor Swift
On August 15, 2024, San Francisco City Attorney David Chiu took the podium at a City Hall press conference. His expression was grim. "This investigation took us to some of the darkest corners of the internet. I am horrified thinking about the women and girls who had to endure this exploitation."
What he announced was the first class action lawsuit in the United States against deepfake pornography websites. The case was People of the State of California v. Sol Ecom, Inc. The defendants were operators of 16 websites. These sites had recorded over 200 million visits in the first half of 2024 alone.
What these websites did was straightforward.
A user uploads a clothed photo of a woman, and the AI 'strips' the clothing from the image. Technically, the system analyzes the original image and synthesizes nude image patterns the AI has learned. The results are nearly indistinguishable from real photographs.
One website's advertising copy read: "Don't waste time on dates. Use our site and you can see her nude right away."
Victims ranged from Hollywood stars to middle school students.
In February 2024, at a middle school in California, AI-generated nude images of 16 eighth-grade girls spread among students. The FBI warned that blackmail cases using AI-generated pornography were surging.
The defendants named in the complaint were scattered across the globe. Sol Ecom, Inc., headquartered in Florida. Itai Tech Ltd. in the United Kingdom. Defirex OÜ and CodeBionic Labs OÜ in Estonia. Augustin Gribinets, a resident of Estonia. And 50 unidentified "John Does." City Attorney Chiu's strategy was to invoke California's Unfair Competition Law. This statute grants the city attorney broad civil litigation authority for consumer protection. It was the same legal tool previously used against opioid pharmaceutical companies, firearm manufacturers, and fossil fuel corporations.
In March 2025, prosecutors filed an amended complaint revealing additional defendants' identities. Richard Teng, Gaofan Xu. And Briver LLC, which operated the sites Undresser.ai and Porngen.art.
The first result came on May 30, 2025. Briver LLC agreed to pay $100,000 in civil penalties and consented to a permanent injunction barring it from ever operating non-consensual deepfake pornography sites. As of June 2, 2025, ten websites had been blocked or shut down in California.
Some defendants fought back, however. Richard Teng, an Illinois resident, raised Section 230 of the Communications Decency Act as his shield. His argument went like this: his website was merely an "intermediary" for third-party algorithms, and a platform cannot be held responsible for content uploaded by users.
There was movement at the federal level as well. The TAKE IT DOWN Act.
The seeds of this legislation were planted in Aledo, Texas, in 2023. A high school student used AI to manipulate ordinary photos of female classmates into nude images and posted them anonymously on Snapchat. One of the victims, Elliston Berry, became the face of the legislative campaign that followed.
An even bigger incident erupted in January 2024. AI-generated pornographic images of pop star Taylor Swift spread across social media. Before they were removed, the images had been viewed over 47 million times. Public outrage boiled over.
Texas Republican Senator Ted Cruz and Democratic Senator Amy Klobuchar introduced the bill on a bipartisan basis. On April 28, 2025, it passed the House 409 to 2. It was near-unanimous. More than 100 organizations, including Meta, TikTok, Google, and Microsoft, expressed their support.
On May 19, 2025, President Donald Trump signed the bill in the White House Rose Garden. Beside him stood First Lady Melania Trump. It was an extension of her "Be Best" anti-cyberbullying campaign. Elliston Berry and another victim, Francesca Mani, also attended the signing ceremony.
The TAKE IT DOWN Act has two core provisions.
First, it designates the "intentional posting" of non-consensual intimate images as a federal crime. The penalty is up to two years in prison for offenses against adults, and up to three years for offenses involving minors.
Second, it imposes a 48-hour removal obligation on platforms. When a victim files a report, the platform must remove the content within 48 hours. Violations are subject to sanctions by the Federal Trade Commission (FTC).
But criticism followed. More than 100 organizations, including the Electronic Frontier Foundation (EFF), the Center for Democracy and Technology (CDT), and the Authors Guild, raised concerns. They argued that the bill's language was vague enough to result in the removal of lawful content. There were worries that the tight 48-hour deadline would lead platforms to delete content without verification. Critics also pointed out that, much like the abuse of DMCA takedown notices, bad-faith reporters could weaponize this law to remove legitimate content.
The TAKE IT DOWN Act became the first U.S. federal law to substantively regulate AI-generated content. Platforms must build reporting systems by May 19, 2026. How many victims will emerge in the meantime, no one knows.
Deepfake technology 'creates' victims' bodies without their consent. The law has only just begun to catch up with the technology. But the disinformation that AI produces is not limited to body images. It includes lies that destroy someone's reputation.
B. Deepfake Financial Fraud
One day in January 2024, a finance department employee at a multinational corporation in Hong Kong received an email.
The sender was the Chief Financial Officer at the company's UK headquarters. The subject was a "secret deal."
The employee was suspicious at first. It looked like a phishing email.
But soon a video conference invitation arrived.
When he logged in, familiar faces filled the screen. The CFO. Several colleagues. They discussed the details of the secret deal.
The employee relaxed. People he knew were speaking to him directly.
He executed 15 wire transfers. To five Hong Kong bank accounts. The total was HK$200 million, roughly US$25.6 million.
When he later checked with UK headquarters, the truth emerged. Every person in that video conference had been fake. They were AI-generated deepfakes.
The victim company was Arup, the 78-year-old British engineering and design firm. They designed the Sydney Opera House. The Bird's Nest stadium for the 2008 Beijing Olympics was their work too. A global company with 18,500 employees across 34 offices worldwide.
Hong Kong police disclosed the case at a February 2024 press conference. Senior Superintendent Baron Chan Shun-ching said: "In a (multi-person) video conference, everyone he saw was fake." Police did not identify the company at the time. In May, Arup confirmed it directly. Rob Greig, Arup's Chief Information Officer, explained: "This was not a traditional cyberattack. Our systems were not compromised and no data was affected." He called it "technology-enabled social engineering."
How did the fraudsters clone the faces and voices of Arup executives?
The answer lay in publicly available materials. Online conferences and company meeting videos. They fed video and audio files already on the internet into AI systems, generating deepfakes in real time.
Michael Kwok, Arup's East Asia regional chairman, warned in an internal memo: "The frequency and sophistication of these attacks is rising sharply worldwide. We all need to learn how to recognize the various techniques that fraudsters use."
The Arup case was the tip of the iceberg. At the same press conference, Hong Kong police announced six arrests related to deepfake fraud. Between July and September 2023, eight stolen Hong Kong identity cards had been used for 90 loan applications and 54 bank account registrations. AI deepfakes were used to fool facial recognition programs at least 20 times.
The use of deepfakes in financial fraud is surging. Traditional phishing emails are easy to doubt. But when you see and hear a familiar colleague, the story changes. Humans evolved to trust visual and auditory information. Deepfakes exploit exactly that instinct.
Voice phishing attacks targeting Citibank have also been reported. The New York State Attorney General filed related lawsuits. The FBI warned that blackmail schemes using AI-generated non-consensual intimate images are increasing rapidly.
Business Email Compromise (BEC) is evolving too. In the past, attackers simply forged email addresses. Now they clone voices for phone calls and use deepfakes for video conferences. In 2024, advertising agency WPP faced a similar attempt. The fraudsters created a WhatsApp account, set up a Microsoft Teams meeting, edited executive footage sourced from YouTube, and cloned the executive's voice. Fortunately, an employee's suspicion foiled the scheme.
Greig said in an interview with the World Economic Forum: "This happens far more often than people think. Visual and audio cues are extremely important to humans, and these technologies exploit precisely that. We really need to start questioning what we see."
Arup stated that its financial stability and business operations were unaffected. The $25.6 million was not fatal for them. But most companies would not be so fortunate.
As deepfake technology grows more sophisticated, traditional identity verification methods become powerless. How do you prove in a video conference that "you are really you"? This question extends beyond financial fraud, reaching into the realm of courtroom evidence.
C. Evidentiary Challenges
In July 2024, a YouTuber named Christopher Kohls uploaded a video. He was a political satire content creator who went by the name 'Mr. Reagan.'
The video featured Vice President Kamala Harris. In her voice, it mocked her presidential candidacy. It was, of course, a fake voice generated by AI.
Elon Musk retweeted the video. It surpassed 100 million views.
On September 17, 2024, California Governor Gavin Newsom signed AB 2839.
Its official title was "Elections: Deceptive Media in Advertisements." The law prohibited the distribution of "materially deceptive content" from 120 days before an election through 60 days after. It granted a private right of action allowing anyone to seek damages.
The law included an exception for satire and parody. But there was a condition. The content had to include the statement: "This [image/audio/video] has been manipulated for purposes of satire or parody." The disclaimer had to appear for the entire duration of the video, in the largest font size used.
Kohls filed a lawsuit the next day. The defendants were Attorney General Rob Bonta and Secretary of State Shirley Weber.
On October 2, 2024, Judge John A. Mendez of the U.S. District Court for the Eastern District of California issued a preliminary injunction. His ruling: AB 2839 was unconstitutional.
Judge Mendez wrote: "Most of AB 2839 acts as a hammer instead of a scalpel, serving as a blunt tool that hinders humorous expression and unconstitutionally chills the free and unfettered exchange of ideas which is essential to American democratic debate." The issue was the First Amendment. The U.S. Constitution broadly protects freedom of speech. Content-based speech regulations must pass strict scrutiny. The government must prove it has a "compelling interest" and that the regulation is the "least restrictive means" of achieving it.
The court found that AB 2839 failed to meet this standard. It noted that the disclaimer requirement "drowns out the message that a parody or satirical video is trying to convey." This was not a "narrowly tailored" regulation.
The judge also found the law's scope excessively broad. "The legitimate sweep of AB 2839 is trivial compared to the substantial number of plainly unconstitutional applications, such as the one in this case."
First Amendment experts had urged Governor Newsom to veto the bill. David Loy, legal director of the First Amendment Coalition, said: "If something is truly defamatory, there is an entire body of law and well-established legal standards for proving a defamation claim. The government does not have the freedom to create new categories of speech outside the First Amendment."
AB 2839 was not the only law challenged. AB 2655, the "Defending Democracy from Deepfake Deception Act of 2024," enacted around the same time, met a similar fate. That law required large online platforms to block or label "materially deceptive content." Elon Musk's X filed a lawsuit against California. On August 5, 2024, Judge Mendez ruled that law unconstitutional as well.
As of 2024, 26 states have enacted laws regulating political deepfakes. Their approaches fall into two broad categories. Minnesota and Texas prohibit the release of political deepfakes during a set period before elections. The remaining 24 states adopt a disclosure-based approach. What about when deepfakes are used as evidence? California SB 970 requires deepfake service providers to issue warnings. In Matter of Weber, reliability concerns about AI calculations affected the admissibility of evidence.
A more fundamental question exists. In the deepfake era, how can we trust video evidence? If any video could have been manipulated by AI, what happens to the evidentiary value of video in court?
This question affects both civil litigation and criminal trials.
When prosecutors submit CCTV footage as evidence, the defense can argue: "How can you prove this is not a deepfake?" Conversely, when a defendant submits video proving an alibi, prosecutors can make the same argument.
Technology moves faster than law. Creating deepfakes is getting easier. Detecting them is getting harder. How will the law bridge this gap? That question leads us into the territory of biometric data and facial recognition surveillance, which we will address in the next chapter.
Kim Kyung-jin
Attorney · Former Member of the National Assembly · AI Policy Researcher
© 2026 Kim Kyung-jin. All rights reserved.