The answer did not come from Ginsburg’s many court opinions, but from a late AI model of justice released on Tuesday. Good or bad, the RBG robot concluded, “it’s ironed out, so it’s none of my business to think about it.”
The form, called Ask Ruth Bader Ginsburg, is based on Ginsburg’s 27 years of Supreme Court legal writing, along with a combination of news interviews and public speeches. A team from an Israeli artificial intelligence company called AI21 Labs has inserted that log into a complex language processing program, giving the AI the ability, they say, to predict how Ginsburg will respond to questions.
“We wanted to honor a great thinker and leader who has an enjoyable digital experience,” the company says on the AI app’s website. “It’s important to remember that artificial intelligence in general, and language models specifically, still have limitations.”
The tool comes during an intense debate about the ethics of creating technology that mimics human life, particularly when the people involved are not around to provide input. But the creators argue their invention It is a useful and easy-to-use tool to help ordinary people, who may not know much about technology, understand how the field of artificial intelligence is progressing.
“There are not many places where the general public can go and play with real AI,” said Yoav Shoham, co-founder of AI21 Labs. “But you can now.”
In recent years, research labs and companies around the world have raced to build technology that replicates or exceeds human intelligence, It offers ways for people to examine and interact with their work along the way.
OpenAI, an artificial intelligence company backed by Elon Musk, has unveiled a text generator, called GPT-3 that can write movie scripts and boost the image generator, DALL-E 2, which translates text commands into creative, and sometimes psychedelic images. In 2020, Shoham created Wordtune, a tool that suggests different ways to write sentences. They followed the release a year later with Wordtune Read, which summarizes the main points of the long, dense passages.
But with the improvement of AI technology, Shoham He said many areas around the field divided. It brings out people of all kinds [thoughts] On … automation that has nothing to do with reality.” “I don’t want people to be disappointed by the poor performance of the current AI and I don’t want them to provoke fear.”
The general public needs to make their own decisions, he said, and his team’s RBG team model is an accessible, practical way to interact with technology.
To build it, the researchers used Jurassic-1, a neural network they created that analyzes a large set of data and develops its own language to output the results to questions or prompts. Neural networks are a computer architecture that attempts to imitate the human brain by processing information.
They fed the model with nearly 600,000 Ginsburg words and created a tool that allows anyone to ask questions of it, which provides answers based on a massive body of writing. “It gives you access to the kind of wisdom that someone who we hold in the most respect has,” Shoham said.
Paul Schiff Berman, a law professor at George Washington University who worked in Ginsburg from 1997 to 1998, said that when he saw the robot, he was amused.
Immediately, he tried to ask her a question that would have been interested in obtaining Ginsburg’s opinion on: “Should federal courts take into account the factual findings of state courts?”
Berman says the response left much to be desired. The model did not directly answer the question and implied that Ginsburg did not believe in the judicial concept of respect, which is incorrect, he said. Berman also noted that the model did a poor job of replicating her unique speaking and writing style.
“I used to think that was something that AI could have imitated better,” he said. “If this is the best [technology] We can do it, we still have ways to go.”
Meanwhile, many AI tech experts have raised concerns about the experiment.
Emily Bender, a professor of linguistics at the University of Washington, said she understands that the experiment’s creators come from a place where Ginsburg is respected, but a hint of technology can think and reason like delayed justice isn’t subtle. “She can utter the words,” she said, “and the style of those words will be informed by the style of the text you feed him, but he does no logic.”
Bender added that linguistic research shows that when people encounter “coherent texts” on a topic they care about, there is a risk that they will take it seriously when they don’t. “People might use this to make arguments in the world and say, ‘Well, maybe RBG said, ‘This AI’ [model] Tell me that.”
The robot is involved but should not be confused with actual legal advice, said Meredith Broussard, associate professor and artificial intelligence researcher at New York University. “It’s really fun to play,” she said, “but we shouldn’t take it too seriously and we shouldn’t pretend this is a lawyer.” (AI 21 states that the model is “just an experiment” and that it can give inaccurate responses that should be taken “with caution.”)
Broussard added that the technology seems no more advanced than ELIZA, a chatbot created by MIT researchers in the 1960s in which a computer program replicated a processor well enough to make people think it was human. She added that there could be a limit to how advanced this type of AI technology can be.
“There is a ceiling to technology because it is not a brain, it is a machine,” she said. “And she just does the math.”
#model #recreate #mind #Ruth #Bader #Ginsburg
