What is Artificial Intelligence
AI is a machine’s ability to perform the cognitive functions we associate with human minds, such as perceiving, reasoning, learning, interacting with the environment, problem-solving, and even exercising creativity. This is done by taking in a myriad of data, processing it, and learning from the past in order to streamline and improve in the future.[i] You’ve probably interacted with AI even if you don’t realize it—voice assistants like Siri and Alexa are founded on AI technology, as are some customer service chatbots that pop up to help you navigate websites.[ii]
From Calculating Machines to Machine Learning
While Artificial Intelligence is the latest technological development that might disrupt all areas of modern work environments, the idea of machines assisting humans not only in physical but also in in cognitive tasks is not a new one. This concept traces back to the works of visionaries like the German inventor and mathematician Gottfried Leibniz (in fact, he was a so-called “universal scholar”), who in the late 17th century invented the calculating machine, a primitive precursor to modern computers. Leibniz’s innovation was driven by the idea that machines could extend the capacity of the human mind, a notion that is central to contemporary AI.
As AI continues to evolve, it embodies this vision of mechanical devices that augment human reasoning, demonstrating the enduring influence of his early technological pursuits on the sophisticated computational systems that today tackle complex problem-solving tasks across various industries.
From Artificial Intelligence to Generative AI
Generative AI is a type of artificial intelligence technology that can produce various types of content, including text, imagery, audio and synthetic data. Generative AI starts with a prompt in the form of a text, an image, a video, a design, or any input that the AI system can process. Various AI algorithms then return new content in response to the prompt. Content can include essays, solutions to problems, or realistic fakes created from pictures or audio of a person.[iii]
Generative AI in Litigation
Law firms and companies have started creating their own proprietary tools to customize generative AI tools for internal use to increase the accuracy of generative AI’s responses and protect confidential information. By providing the proper prompts and sufficient supporting documents or information, generative AI can help attorneys perform certain types of legal tasks more efficiently.[iv]
Potential Uses & Benefits
Litigators may use generative AI to:
- Produce initial drafts of correspondence.
- Create litigation timelines with dates, locations, and key actors based on information collected, for example, from pre-litigation investigation, pleadings, and discovery.
- Produce initial drafts of pleadings.
- Review and synthesize voluminous discovery.
- Gather information on witnesses and create outlines for depositions.
- Provide a starting point for legal research and produce initial drafts of briefs.
- Produce initial drafts of opening statements and closing arguments.
- Review and summarize trial exhibits.
Key Risks & Limitations
While generative AI has many benefits, it also comes with known risks and limitations that counsel must consider, including hallucinations, bias, and confidentiality issues. Additionally, counsel should keep in mind that AI-generated evidence may pose discovery and evidentiary issues, including those related to admissibility in court. [v]
Generative AI to Deepfakes
As outlined above, thanks to advances in artificial intelligence, it is now possible to create a genuine-looking video that makes real people appear to do and say things they never did or said. These “Deepfake” videos, images, and audio stand poised to add a new chapter to trial courts’ history of dealing with inauthentic evidence. “Deepfake” software programs use artificial intelligence (AI) to produce forged videos of real (and even entirely fabricated) people that appear genuine, making them appear to do and say things they never did or said. The more video and audio footage of real people that can be fed into the system, the more convincing the result. Software for creating deepfakes is already freely available online and fairly easy for anyone to use. As the software’s usability and the videos’ verisimilitude keep improving over time, it will become harder for laypeople, as well as computer systems, to tell real from fake.”[vi]
Deepfakes are still a nascent topic in the law. While deepfake-related research in computer science dates back to at least mid-2014, it took a few years for the technology to get to the point where it began to concern legal scholars and policymakers. So far, most of their efforts have focused on what we might call the question of containment: how to prevent, mitigate, and punish the abuse of deepfake technology for harmful purposes. With that said, it is safe to assume that the ready availability of deepfake tools and antisocial uses thereof will continue, irrespective of how the law may attempt to contain, regulate, and punish them. If deepfakes are here to stay, then we must be ready to respond to their effects on our legal system. [vii]
Deepfakes in the Courtroom
Deepfakes may arise in the evidentiary context in several ways. A party might fabricate a video specifically for purposes of the litigation in order to try to prevail. Or a litigant might encounter a deepfake video made by someone else and wish to introduce it into evidence, not realizing it is fake. Fake videos may end up (whether accidentally or maliciously) in archives that have historically been considered trustworthy, such as those of news outlets. Suppose their presence goes undetected by the custodian of those records. In that case, there is a risk that the custodian might unwittingly vouch for a deepfake when called upon to authenticate evidence in a court proceeding.[viii]
Moreover, even in cases that do not involve fake videos, the very existence of deepfakes will complicate the task of authenticating real evidence. The opponent of an authentic video may allege that it is a deepfake in order to try to exclude it from evidence or at least sow doubt in the jury’s minds. Eventually, courts may see a “reverse CSI effect” among jurors. In the age of deepfakes, jurors may start expecting the proponent of a video to use sophisticated technology to prove to their satisfaction that the video is not fake. More broadly, if juries—entrusted with the crucial role of finders of fact—start to doubt that it is possible to know what is real, their skepticism could undermine the justice system as a whole. [ix]
Deepfakes’ Ramifications for Genuine Evidence
Deepfakes’ authentication difficulties are twofold. One problem is how to show a photograph, video, or audio is fake. The other is how to show it isn’t. As deepfakes become increasingly common and realistic, their very existence will undermine the reliability of genuine evidence, creating issues for the proponents of authentic photographs, videos, and audio recordings.
In the era of digital evidence, practitioners are already accustomed to bringing and responding to courtroom challenges of digital photo, video, and sound recordings, despite Rule 901’s relatively low standard for authentication. However, as real and fake become harder to distinguish, such challenges may be harder for the proponent to overcome. Successfully getting a video admitted into evidence may require additional motion practice, witness testimony, and forensic tools. These hurdles threaten to exclude genuine evidence from being admitted altogether.[x]
[i] What is the history of Artificial Intelligence (AI)? Available at https://www.tableau.com/data-insights/ai/history#:~:text=Birth%20of%20AI%3A%201950%2D1956&text=The%20term%20“artificial%20intelligence”%20was,intelligence%20called%20The%20Imitation%20Game.
[ii] What is AI (Artificial Intelligence)? Available at https://www.mckinsey.com/featured-insights/mckinsey-explainers/what-is-ai
[iii] What is generative AI? Everything you need to know. Available at https://www.techtarget.com/searchenterpriseai/definition/generative-AI
[iv] Generative AI in Litigation. Available at https://www.reuters.com/practical-law-the-journal/litigation/generative-ai-litigation-2024-03-01/#:~:text=Trial%20and%20Evidence,for%20use%20in%20the%20courtroom
[v] Id.
[vi] Deepfakes In The Courtroom. Available at https://deliverypdf.ssrn.com/delivery.php?ID=742110126119065022097064104099020127062041020052054087069121081029067094031114096088102031125025020116062122090076095120116081057017017032053103083073117126014121087041058031020110000081098118072015000085099022091124123030122010091104001121122125103097&EXT=pdf&INDEX=TRUE
[vii] Id.
[viii] Id.
[ix] Id.
[x] Id.