“Artificial intelligence (AI) is like a highly skilled psychopath,” said U.S. crime expert Dr. Brent Turvey (54) in an interview with Chosun Ilbo on Oct. 23. He explained, “AI’s only goal is to produce the most efficient result based on data.” Turvey, a profiler and crime scene analyst since 1996, is the author of “Criminal Profiling” (1999), a key text in the field. His analyses are widely regarded as credible evidence in U.S. federal and state courts.
Turvey attended as a speaker at the 10th International CSI Conference (CSI Korea 2024), held in Incheon from Oct. 23 to 25 and hosted by the Korean National Police Agency. In his lecture and interview on the interaction between human intelligence (HI) and AI, he stated, “Human analysts constantly re-evaluate their conclusions, weighing the ethical consequences of their decisions. In contrast, AI purely calculates without any moral consideration,” warning that AI’s ability to manipulate disinformation poses a significant risk to the criminal justice system.
“When fabricated evidence is submitted in court, it can greatly affect a judge’s ruling,” he said, citing the 2018 defamation lawsuit between Hollywood stars Johnny Depp and Amber Heard as an example. Heard claimed Depp had abused her and submitted photos of her bruises as evidence. However, Depp’s team countered that the bruises were exaggerated using photo editing software, and Depp won the case in June 2022.
Turvey also highlighted instances where lawyers introduced “false precedents” generated by AI during trials. He referred to a case involving Michael Cohen, former personal attorney to Donald Trump, where AI-generated favorable precedents were cited. However, the U.S. District Court for the Southern District of New York rejected the statement, noting, “None of the cited cases could be found.” This process also revealed that AI-generated content had been used.
The U.S. crime expert emphasized that human “laziness” and “carelessness” make people highly vulnerable to disinformation generated by AI. He explained, “The Michael Cohen case illustrates how many people use AI-generated content without thoroughly reviewing or verifying it, leading to absurd situations.” He noted, “It’s not uncommon to see research papers with comments like, ‘I am an AI language model and do not have real-time access to information,’ or ‘This is a suggested introduction for your topic’ copied and pasted word-for-word,” adding that “This is a clear result of relying too heavily on generative AI without even reading the output.”
Turvey also pointed out that AI-generated deepfake photos and videos pose a significant threat to the upcoming U.S. elections. In August, a video featuring global pop star Taylor Swift, falsely showing her endorsing Donald Trump, went viral on social media. However, she had already expressed her support for Democratic candidate Kamala Harris. Turvey noted, “Given that Swift has around 300 million followers on social media, it’s possible that someone created this deceptive deepfake video to exploit her influence.”
He expressed concern that AI could become a powerful tool that learns and reproduces existing discrimination in human society. As the saying in the data industry goes, “Garbage in, garbage out,” meaning AI lacks the ability to make value judgments about data. Turvey cited facial recognition AI as an example. “Since the system is primarily trained on facial data of white individuals, it fails to accurately extract and analyze features of other races,” he explained. “This leads to an AI profiling system that is effective only for white individuals, which can ultimately undermine the reliability of investigative results.”
