In the realm of investigations, the utilization of artificial intelligence has become an increasingly attractive tool for attorneys seeking efficient ways to uncover information about individuals and organizations. However, the accuracy of AI-generated results can be questionable at times, leading to outcomes that seem more akin to fictional narratives than reliable facts. Bruce Gertsman, President of Waterfront Intelligence, an investigations consultancy unearthing evidence, has conducted numerous investigations with the assistance of AI, and recently provided his views on the complexities associated therewith.
Exploring the Potential of ChatGPT
As a journalist engaged in investigative work, Bruce decided to put ChatGPT to the test, delving into its capabilities and limitations. On one hand, this tool offers attorneys a valuable pathway to explore leads, providing insightful suggestions on areas worth investigating. On the other hand, the AI-driven outcomes occasionally present glaring inaccuracies that could easily be mistaken for concocted stories.
Typically, an individual under investigation should have some level of media or online presence. By posing questions regarding arrests or lawsuits, ChatGPT can quickly generate a series of instances where the subject has encountered legal entanglements. These instances serve as starting points for further exploration, highlighting specific disputes or legal troubles.
In one particular case, Bruce’s focus centered on a hedge fund owner. Initial inquiries about his background yielded concise and accurate summaries sourced from online biographies. However, these summaries lacked the crucial details he sought. Digging deeper, Bruce asked ChatGPT about potential legal battles, criminal charges, and adversaries. To his surprise, ChatGPT provided information on four distinct lawsuits with corresponding allegations, all related to defamation. These facts turned out to be accurate, providing him with a clear direction for additional research.
Armed with these foundational pieces, an attorney can leverage existing knowledge and employ the services of an investigator to delve deeper into the specifics. While not comprehensive, this approach acts as a guide, saving both time and resources during a more extensive investigation.
A Shocking Revelation
At the culmination of an investigation involving an executive, Bruce decided to put ChatGPT to the ultimate test: Could it uncover information he had missed? He queried the tool about any lawsuits the executive had faced in the past. The response provided astonishing details, stating that the U.S. Securities and Exchange Commission (SEC) had sued the executive and their company, listing a specific date, court, and settlement amount.
However, upon thorough examination of government databases and archived media reports, no evidence of such a case was found. Even contacting the SEC and the Judge’s clerk proved futile in confirming the existence of this alleged lawsuit. It became apparent that ChatGPT had fabricated a case that had never existed in reality.
This incident serves as a cautionary tale, underscoring the potential dangers of disseminating AI-generated information to a wider audience. Had this erroneous answer been circulated, it could have sparked baseless rumors, inflicting irreparable harm to the executive’s reputation. As attorneys piece together fragments of an individual’s background, it is essential to approach the findings with caution, as a comprehensive investigation often uncovers inaccuracies lurking within supposedly factual details.
The case of ChatGPT exemplifies the intricate challenges that arise when AI intersects with the investigative process. While it undoubtedly provides valuable leads, the need for meticulous fact-checking and human validation remains paramount. As the utilization of AI in investigations becomes more prevalent, striking the delicate balance between efficiency and accuracy will be essential to uphold the integrity of our pursuit of truth.
