The Week in Review: June 9, 2023

It was bound to happen at some point. After using ChatGPT irresponsibly, a lawyer could face sanctions, pending a judge’s review.

Though OpenAI itself warns users that ChatGPT may generate false and completely made-up information, many folks are pushing the limits to see just how much they can get away with.

It looks like Manhattan lawyer Stephen Schwartz just found out how much he can get away with.

Read up on that and other language tech stories (like how some AI researchers are reviving ancient speech) below.

Got news you’d like to share in next week’s Week in Review? Send it over to our editorial team here.

Lawyer Who Used ChatGPT Faces Penalty for Made Up Citations (via The New York Times)

This week, a Manhattan lawyer found himself in hot water as a judge probed him on his use of ChatGPT to generate a legal briefing chock-full of completely made-up citations and judge’s opinions.

When pressed on his use of the chatbot, Stephen Schwartz told a judge that he hadn’t considered the idea that ChatGPT would “fabricate” case information. The judge is now considering whether to implement sanctions on Schwartz and his partner, who was also named on the brief.

“I continued to be duped by ChatGPT. It’s embarrassing,” Schwartz said in the hearing.

Nvidia invests in Google-linked generative A.I. startup Cohere (via CNBC)

Cohere — a startup founded by Ivan Zhang and Google alumni Aidan Gomez and Nick Frosst — recently raised $270 million in the company’s Series C round of venture capital funding. Among the company’s big backers are Nvidia, Oracle, and Salesforce.

“The team at Cohere has made foundational contributions to generative AI,” said Nvidia’s CEO, Jensen Huang. “Their service will help enterprises around the world harness these capabilities to automate and accelerate.”

How caregiver speech shapes infant brain (via Science Daily)

A study recently published in Developmental Cognitive Neuroscience has found evidence that infants’ brain development improves the more frequently their caregivers speak with them. According to the paper, infants who were exposed to more words experienced slower development of white matter in their brain, indicating a “cognitive advantage.”

“This work highlights parents as change agents in their children’s lives, with the potential to have enormous protective effects,” one of the researchers told Science Daily. “I hope our work empowers parents with the knowledge and skills to support their children as best they can.”

AI has brought back 15 languages people haven’t heard for centuries. Here’s what they sound like. (via Upworthy)

Ever wonder what the Vikings sounded like when they spoke? Now you don’t have to — EquatorAI recently used artificial intelligence to create a series of videos demonstrating what ancient languages might have sounded like back in their heyday. Though it might not be 100% accurate, it’s a fascinating glimpse back in time.