Up Next

AI in the Courts: The Jury Is Out


A session on the role of emerging technologies in the courtroom was part of last month’s New York State Bar Association Annual Meeting in New York City.

“Emerging Technologies in Litigation” included a panel of local and federal judges as well as an e-discovery researcher and emerging technology attorney. The group discussed the use of artificial intelligence in the courtroom.

The session addressed the role that AI could play in judicial decision making, where algorithms potentially can predict behavior and outcomes resulting from different legal strategies. The rationale is that law is based on precedent — if a case is similar to past cases, then the results shouldn’t be all too surprising.

However, given the rise of deepfakes — and the possibility that AI in effect could manufacture evidence — some argued that the technology should be excluded from court proceedings.

Despite such concerns, the global “legaltech” market for AI is expected to grow in the coming years, driven by the trend in major law firms to adopt various legaltech solutions that aim to reduce turnaround time for some legal cases.

AI is used to help with document management systems, e-discovery, e-billing, contract management, and even practice and case management.

AI already has been employed at a lower level in the Los Angeles Superior Court to handle seemingly mundane traffic citations. Visitors to the court’s website can interact with Gina, an AI-powered online avatar, to pay a traffic ticket, register for traffic school, or schedule a court date.

Since being installed in 2016, Gina — which is part of an effort by the LA Superior Court to reduce the backlog of cases — has had more than 200,000 interactions a year, and has reduced traffic court wait times dramatically.

One Step Closer to PreCrime

AI’s predictive algorithms can be used by police departments to strategize about where to send patrols, and facial recognition systems can be used to help identify suspects.

Combined, these approaches sound eerily similar to the Philip K. Dick short story, “The Minority Report,” which became the basis of the Steven Spielberg-directed film Minority Report, in which the police department’s PreCrime unit apprehends criminals based on foreknowledge of criminal activity.

“Courts currently are using AI algorithms to determine the defendant’s ‘risk,’ which can range from the probability that the defendant will commit another crime to whether or not they will appear for their next court date for bail, sentencing and parole decisions,” explained technology inventor/consultant Lon Safko.

Often AI can be wrong — not only in determining where officers should patrol, but also in recommending how criminals should be sentenced. Here is where the Correctional Offender Management Profiling for Alternative Sanctions comes into play. It compares defendant answers to questions as well as personal factors against a nationwide data group and assigns a score, which is used to determine sentencing.

“Recently in Wisconsin, a defendant was found guilty for his participation in a drive-by shooting,” Safko told TechNewsWorld.

“While being booked, the suspect answered several questions that were entered into the AI system COMPAS,” he continued. “The judge gave this defendant a long sentence partially because he was labeled ‘high risk’ by this assessment tool.”

AI in the Courts

At the present time it isn’t clear how widespread the use of AI in the courts will be — in part because the courts at all levels have been quite slow to embrace any new technology. This could be changing, however, as AI can help streamline the courts in ways that could benefit all parties.

“We believe the courts are leading digital transformation in the market, and approximately 90 percent of courts have evolved from traditional court reporting to professional digital court reporting,” said Jacques Botbol, vice president of marketing at software firm Verbit.

“Certain applications of AI are often adopted faster than others — particularly those surrounding the automation of routine tasks and workflows,” he told TechNewsWorld.

“It’s interesting to note that AI is also being utilized through more complex applications, such as utilizing AI to make decisions regarding cases,” added Botbol. “These use cases will be adopted more slowly as there are significant concerns about due process, biases, etc.”

AI Court Reporting

Supporters of AI technology in the courts point to how it can help court reporters do their job better.

“Today, most court reporting firms reject work since they don’t have the necessary workforce to handle it all,” explained Botbol. “AI is helping to fill the gaps that the retiring court reporters and the legacy court reporter market have left,” he noted.

At the same time, “lawyers want to receive materials quickly, and today depositions are getting delayed because of the shortage in the market — with some areas reaching more than 35 percent,” Botbol added.

AI, along with automatic speech recognition (ASR), allows for proceedings to be recorded and processed in a timelier manner.

“There is a backlog of cases that need to be transcribed, yet with AI-based ASR tools these transcripts can be processed at faster turnaround times,” said Botbol. “Instead of relying on court transcriptionists, the courts have multiple court reporting agencies that they can assign the work out to in order to clear their backlog and work more efficiently.”

Judge and/or Jury

No one is expecting that AI will fill the role of judge or jury — at least not in the legal system of the United States. However, AI could help ensure that the accused in criminal cases truly are granted the right to a speedy trial, while also addressing the backlogs in the civil courts.

“In the future, AI will not only serve as an add-on, but will also help to streamline trials by removing delays, which will lead to smarter and faster decisions being made,” said Tony Sirna, legal specialist at Verbit.

“Applications of AI are being studied and piloted for a number of use cases,” he told TechNewsWorld.

These include not only sentencing and risk assessment such as COMPAS, but also settlement of disputes.

“Online Dispute Resolution is another aspect where we may see automated adjudication of small civil cases,” noted Sirna.

AI could help the parties reach an equitable settlement in civil cases.

“Mining extensive amounts of related court cases and decisions will come into play, with parties submitting their cases and using AI combined with data mining for settlement options or fair adjudication,” noted Sirna.

AI Rights

Another consideration that likely will come up is how AI will be treated by the courts. Can AI be an “expert witness,” for example? If so, how will AI need to be treated by the courts? Will AI need to be granted some form of rights?

“AI likely won’t need ‘rights,’ but it will need control, and a team that manages the innovation in each court,” said Sirna.

“The aspect of ‘rights’ related to AI poses interesting legal questions: Who is responsible for the AI? Is the AI algorithm fair or biased? At what point does the AI make its own decisions? Who is liable for results or decisions rendered by algorithms — the user, the designer, or the court?” pondered Sirna.

However, many of these questions likely won’t need to be addressed anytime soon — nor will AI have the power to pass judgment.

“Our judicial system is by no means ‘early adopters,’ but for good cause,” said Safko.

“Rendering a just verdict and sentence is paramount, and we have to be sure that the defendants and plaintiffs are properly represented and that their information is protected,” he said. “This is why doctors insist on still using fax machines over email, which can easily be hacked.”

Automated Recommendations

AI could have a place in the courtroom, but perhaps only to aid the human lawyers, judge, court reporters and jury. AI shouldn’t replace any of those humans, but aid them in doing their job.

“Once a technology has proven itself to be reliable and show a time or cost savings, it has been and will be adopted,” suggested Safko.

“AI is not a perfect science — it is still programmed by humans, and not every set of data perfectly matches the predetermined rules programmed into the application,” he warned.

However, with the increasing pressure on court dockets, any time or cost saving measures need to be considered. It is important too, to consider how AI then could affect people’s lives.

“Every automated recommendation should be reviewed by a qualified judge to verify the outcome. Then their recommendation needs to be fed back into that system to allow it to become more proficient at rendering appropriate decisions,” said Safko.

“We can’t risk peoples’ lives on automated apps that save money,” he noted.

“Even the Chief Justice of our Supreme Court, John Roberts, is concerned about how AI is affecting the U.S. legal system,” Safko explained. “When asked about AI in our legal system, he said ‘it’s a day that’s here, and it’s putting a significant strain on how the judiciary goes about doing things.'”

Up Next

Legal Artificial Intelligence in and out of the Courtroom

Artificial Intelligence can conjure both mystery and hesitation. The application of AI and other technology in the legal community is often grounded in more day-to-day concerns. These include automating mundane tasks, reducing costs, accelerating litigation through e-discovery and responding to the need for more access and efficiency. Yet AI-based technologies are accelerating to further aid the legal industry. Regardless of the perceptions of AI, it’s here to stay.

AI in Law

Legal artificial intelligence is centered around the utilization of tools like Natural Language Processing and data analytics to sift through massive quantities of unstructured documents and structured data to speed discovery and uncover links more efficiently and at a more effective cost. Other leading concepts include using data to aid attorneys in jury selection and providing insights, including mining of social media.

Automation of common processes is underway – from software that helps to expedite drafts of common contracts to analyzing billing and costs. Analytics tools are being deployed to develop cost models, determine best practices for staffing and provide decision points on whether certain cases should be settled.

Additionally, law firms are using AI to understand the resources required to handle different kinds of client matters. Clients are increasingly requesting fixed-fee engagements or alternative fee agreements from law firms. But if those firms do not understand their costs, a fixed-fee engagement poses a serious risk of cost overruns borne by the firm.”

AI in the Courts

Legal technology is also improving access to courts and the delivery of legal aid. Code for America, a nonprofit organization, is focused on delivering technology solutions that make government services effective and easy to use. They developed an application called Clear My Record which helps individuals who are eligible to expunge their criminal records. Their mission is that once a record clearance bill passes, governments have the infrastructure in place to automatically deliver relief to people—in a faster and more dignified way.

Other court-based AI initiatives include assisting Online Dispute Resolution, where data models and algorithms compare judgments and cases to produce options for resolution. Although in early stages, and not without concern, the concept is seen in the privately developed app CoParenter. CoParenter’s app and services are designed to aid divorced or separated partners and negotiate decisions about their children.

According to one of the founders, a former family court judge, 80 percent of the disagreements presented in the courtroom didn’t even require legal intervention — but most of the cases included over-involved parents asking the judge to make the co-parenting decision.

AI and Court Reporting

The Wall Street Journal pointed out a very real problem earlier this summer about the attrition in available court reporting resources, especially stenographers. There were 18% fewer stenographers in 2018 than there were three years earlier. Fewer individuals are entering the field, yet the demand for their services continues. The problem will not get better with time, as longtime stenographers continue to retire.

Court reporting and legal transcription are therefore primed to adopt AI and automation. Courts themselves are seeking to reduce costs and make access to court services, such as transcription, more affordable for litigants. There’s also the lack of stenographic services for below-the-line cases (low-value litigation, short depositions) and robust e-discovery to consider.

Many legal proceedings are already being recorded on advanced systems today, which are able to identify speakers. These recordings are then sent to a court reporting agency for transcription and certification. AI combined with Advanced Speech Recognition seeks to accelerate and increase the capacity to produce transcripts. Advanced versions of ASR technologies now incorporate Natural Language Processing (NLP). These systems capture real conversations and use machine intelligence to process them. The accuracy provided by ASR is dependent on many factors, including speaker volume, background noise, the recording equipment used and more.

Learning models or algorithms that are at the core of AI are critical to the adoption of ASR and digital court reporting. As the AI engine processes more terms, more recordings and patterns, the accuracy of the final text output is fine-tuned.

Key Benefits

For court reporting and legal agencies, the application of AI to transcription efforts is paramount. It results in reduced overhead costs, a turnaround of transcripts faster and increased operational capacity to serve more clients.

Benefits to litigators also include access to searchable audio, on-demand transcripts and lower costs of service. Courts systems also benefit from digital recording and transcription for reduced operating costs, as well as further access and faster turnaround times for litigants.

However, technology will not replace the human factor. Combining technology with human-delivered services offers a way to address scalability challenges in the face of a dwindling supply of stenographers, while still ensuring quality. For the highest accuracy, an individual still needs to listen to the recording and correct the technology-captured transcript for the record.

Back To Top