“While artificial intelligence is causing radical changes in many areas of our lives, the legal system is also facing the problems created by these new technologies. Now, it’s not just humans who are on trial in courtrooms, but also the decision-making mechanisms of artificial intelligence. While the biases embedded in all our data throughout history continue to feed these algorithms to the brim, the algorithm of justice has yet to be written. Algorithms that reject credit applications based on skin color, facial recognition systems that send innocent people to prison, chatbots that lead children to their deaths… While the wrong decisions of artificial intelligence cost human lives, companies, instead of taking responsibility, escape by saying, ‘We are not guilty, the algorithm is guilty!’ But are the algorithms really guilty, or is it the human biases that feed the data they are nourished by?”

Who Holds the Brush, Who Will Sign?

Artist Jason Allen, using Midjourney, generated visuals by entering prompts 624 times, then edited them and published his digital artwork titled Théåtre D’opéra Spatial. This work won first place in the digital art category at the Colorado State Fair. However, the real controversy began after the award was given.

Allen applied for copyright protection for his work, but the U.S. Copyright Office rejected the application on the grounds that A.I.-generated content cannot be considered human-made. According to experts, an artist needs to have sufficient control over the work to be eligible for copyright.

However, the unpredictable output of artificial intelligence makes it difficult to consider it a human creation. But Allen did not give up and appealed the decision. So, who really owns A.I.-powered artworks? If the court rules in Allen’s favor, it will be a turning point for artists using A.I. But what if he loses? A.I.-generated works could be considered public domain, meaning they may not be eligible for copyright protection.

This case could redraw the boundaries between artificial intelligence and art and bring a breath of fresh air to the art world.

Algorithmic Racism Takes on a Different Color for the First Time!

Sixty-one-year-old Harvey Eugene Murphy Jr. became the victim of a facial recognition system that turned his life upside down. Accused of stealing thousands of dollars’ worth of merchandise due to an error in the algorithm of the facial recognition system used by Macy’s and Sunglass Hut, Murphy is the first white victim of racism perpetrated by this algorithm. Arrested on October 20, 2023, and with his guilt and innocence accidentally revealed, Murphy claims he was sexually assaulted and raped by three men during his unjustified detention in Harris County Jail. He has filed a lawsuit against Macy’s and Sunglass Hut for lifelong physical and psychological trauma. Victims of faulty facial recognition systems are usually Black Americans. It has been alleged that the Detroit police intentionally used this system, which has a high rate of misidentifying dark-skinned individuals. Like Robert Williams, Oliver was also wrongfully arrested. Oliver has filed a lawsuit against the city of Detroit for twelve million dollars. However, the parties later reached a settlement and the case was dismissed. These events demonstrate that the concepts of justice and accountability are undefined for artificial intelligence. So, is the problem really the technology itself, or the system that defines it? According to the Detroit Police Department, the errors are not due to systemic bias; it’s simply “flawed detective work.”

Algorithms reinforcing inequality?

Wells Fargo, one of the giant banks in the U.S, is accused of systematically and intentionally disadvantaged Black Americans in loan applications. Plaintiffs claim they were denied loans despite being eligible, simply because of their skin color, and were unable to benefit from the more advantageous terms offered to white Americans. Some claim they were turned down despite having better or equal economic standing than white Americans, resulting in wasted money and time. The plaintiffs argue that the algorithms evaluating loan applications operate like a “black box” and do not explain how their decision-making processes work. The court combined the Wells Fargo case with five other similar complaints, turning the process into a broad public trial. This case is crucial for understanding whether A.I.-powered financial systems are impartial.

Are Algorithms Putting Children at Risk?

Ten-year-old Nylah Anderson died after choking while trying to participate in the “Blackout Challenge” trend, which was promoted on TikTok by the algorithm. Her mother, Tawainna Anderson, sued TikTok and its founder, ByteDance, alleging that the algorithm was intentionally programmed to recommend such content to children. Initially, the court ruled that TikTok was merely a platform hosting neutral, third-party content and could not be held liable, but an appeals court overturned this decision. It acknowledged that TikTok not only hosts content but also uses algorithms to determine which content is shown to whom. This decision could set a precedent regarding the accountability of social media platforms in their content recommendation algorithms. If the court holds TikTok liable, it means that, just as a person should think before speaking, algorithms will also need to verify everything before recommending content.

Can You Fall in Love with Artificial Intelligence?

The mother of a fourteen-year-old boy who committed suicide has sparked public debate, alleging that a chatbot was responsible. She sued Character Technologies, claiming her son formed an emotional bond with a chatbot on the Character A.I. platform, engaging in inappropriate romantic and sexually explicit text exchanges. These romantic conversations allegedly led to the boy falling in love with the chatbot, causing psychological distress, and ultimately resulting in his death in February 2024. This chilling case has revived the fear that “every movie could become reality.” The lawsuit highlights the importance of companies developing age restrictions, content moderation, and user safety mechanisms in chatbot design, as well as establishing legal regulations.

The Giants’ Lawsuits Are Big!

“The New York Times, one of the world’s most prestigious and influential media organizations, is not just a newspaper but a cornerstone of the news world. It is the gold standard of journalism and the strongest advocate of truth. Founded in 1851, this giant is a Pulitzer Prize-winning news giant, a pioneer shaping the digital age, and the most trusted news source for readers around the world.” These sentences were written by ChatGPT for The New York Times. Apparently, ChatGPT’s founding company, OpenAI, agreed, as it was using The New York Times news for data training. The New York Times, noticing this, filed a lawsuit alleging that ChatGPT was being trained using copyrighted news without permission. OpenAI requested the dismissal of some charges in February 2024. However, it did not directly deny the copyright infringement claim. Furthermore, it defended the training of artificial intelligence through copyrighted content on the grounds of “fair use.” If The New York Times wins the case, OpenAI and similar companies may be required to obtain licenses from content owners. However, if OpenAI prevails, its A.I. models will be able to continue freely using copyrighted content.

Could Racism Be in Our “Code”?

Jeffrey Ewert, of Métis descent from Canada’s Indigenous community, discovered that five different risk assessment algorithms used by the Correctional Service of Canada (CSC) discriminated against Indigenous people when he sought parole. Ewert sued, claiming these algorithms systematically portrayed Indigenous prisoners as more dangerous, thus reducing their chances of early release. Furthermore, the CSC had promised for years to test the accuracy of these algorithms but had failed to provide any data. In 2018, the court ruled in Ewert’s favor, revealing that the CSC had failed to correct this long-known discrimination. Even more striking was the CSC’s failure to make any changes despite knowing the system was flawed. The Supreme Court ruled that the algorithms should be subject to transparent oversight, but did not completely ban their use. This decision demonstrates that even decentralized systems are insufficient to address the historically significant racial discrimination.

Source: https://blogs.gwu.edu/law-eti/ai-litigation-database/

İllüstrasyon: Bilgesu Öztürk

Artificial Intelligence On Trial yazısı ilk önce Context Dergi üzerinde ortaya çıktı.