By Sejal Govindarao
PHOENIX (AP) — The Arizona Supreme Court has created a pair of AI-generated avatars to provide news of all the judgments issued by the judicial authorities.
While Florida courts use animated chatbots to help visitors navigate the website, the Arizona Supreme Court is diagramming the creation of Victoria and Daniel and the new territory. The two avatars, made of pixels, do different tasks in that they act as faces for news from the courts, as spokesmen made of flesh and blood, do, but are faster.
The use of AI has touched on almost every occupation and discipline, and has grown exponentially in recent years, showing endless possibilities when it comes to simple things like internet searches and things as complex as brain surgery. For Arizona Supreme Court officials, their ventures for AI are rooted in a desire to promote trust and trust in the judicial system.
What helped solidify the court’s need for more public outreach?
There was a protest outside the Capitol last April, calling for the two to indict justice after the Arizona Supreme Court ruled that civil war-era laws banning almost all abortions, unless women’s lives are at risk. Emotions erupted on both sides of the problem.
When Secretary Anne Timer took over the court last summer, she made the Public Trust a key pillar of her platform. She had already been thinking about how to use digital media for several years to reach out to the public. The abortion ruling also helped solidify the idea that, among other judgments, the courts need to be part of the story in order for people to learn about their opinions and their meaning.
“We serve the public a lot by saying we have issued this decision,” she said. “Now, let’s help you understand what it is.”
Timmer told The Associated Press earlier this year that if the court had to re-authorize abortion, it had to approach information dissemination in a different way. In an interview Wednesday, she said news releases and avatar videos could help the public better understand the legal foundations of long decisions.
“We got a lot of backlash for it, but perhaps obviously, in that people can complain that people don’t understand what we did when they weren’t doing enough to give a simplified version, she explained in an interview in January that she wanted to know the basis for a court decision and what they could do, such as lobbing the state’s lobbyk to help change states for what they could.
Democratic government official Katie Hobbs signed the repeal of the ban last May, and in November Arizona voters approved a constitutional amendment to expand abortion access to fetal survival.
Who are Daniel and Victoria and how do they work?
It will deliver court news releases in a way created with the programs called Creatify, Daniel, and Victoria. Videos featuring either one have been posted for all judgments by the High Court and may be used to access future judicial projects, community programs and citizen information.
The court has been sending out releases since October to summarise and explain the award. After a successful release, we began exploring options to convey that information through video.
Avatars generated by AI were the most efficient way to create videos and publish information, court spokesman Alberto Rodriguez said. He said that video production can usually take several hours, but the video generated by AI is ready in about 30 minutes. The court may introduce more AI-generated reporters in the future, Rodriguez said in a news release.
Justice writing legal opinions must also draft a news release, and the language must be approved across the bench. The judiciary then works with the court’s communications team to script the avatar. The avatar did not interpret the original court’s decision or opinion, Rodriguez said.
Daniel and Victoria’s names and physical appearance were designed to represent a wider cross-section of people, Rodriguez said. He said they were not going to come out as real people, and the court said in its disclaimer the origins of their AI. The court is exploring a variety of emotional delivery, rhythm, pronunciation and Spanish translations, as well as Spanish translations for avatars, Rodriguez said.
Does an avatar resonate with the audience?
Mason Colts, a clinical instructor at the Harvard Law School Cyber Law Clinic at Berkman Klein Center for Internet and Society, described the new cyber employees at the court as “very realistic.” Their voice may be given to them, but he said that if viewers read subtitles and watch the character’s movements and expressions, he could be fooled into thinking Daniel and Victoria are real reporters.
Kortz also said it would be better to see the disclaimer language in the text description of the video more prominently featured.
“You want to make it as difficult as possible for someone to spread the disclaimer or inadvertently remove it,” he said.
Ashley Landrum, an associate professor at Walter Cronkite Journalism and Mass Communications School at Arizona State University, said that the avatar feels a robot. She said that forms of mimicking authentic dialogue and storytelling may be more appealing than AI reading in news releases.
“It’s not just about using AI or creating videos,” she said.
Still, it’s a great line. She said that attractive traits help build trust over time, but the danger means that content can appear biased.
Original issue: March 18, 2025, 1:19pm
