In a busy 48 hours, OpenAI could be in more legal trouble and its ChatGPT chatbot has been accused of a significant leftwing bias, overshadowing the company’s first public acquisition in its seven-year history.
The New York Times is reported to be weighing whether to sue the maker of ChatGPT over claims that the AI-powered bot has infringed its intellectual property rights by using articles from the newspaper to train and power its artificial intelligence system without permission or compensation.
The Times and OpenAI have been in negotiations for weeks over reaching a licensing agreement, but the talks have stalled and turned sour, according a report by NPR.
A potential lawsuit could ensue, the report said, which could have far-reaching implications for the future of generative AI and journalism.
ChatGPT and other similar AI chatbots powered by large language models that have been trained on a massive amount of data scraped from the internet, including millions of articles from The Times and other news outlets.
Last month, OpenAI and Meta Platforms were sued by a group of authors over alleged copyright infringement.
Comedian Sarah Silverman and fellow authors Christopher Golden and Richard Kadrey launched the lawsuit with a claim that their work was used without their permission to train the chatbots.
If that is true, further news today that ChatGPT has been accused of systematic political bias toward the US Democrats, centre-left Lula in Brazil and the centre-left Labour Party in the UK, is possibly less surprising.
While Kadrey has said he “hate[s] politics. It’s the lowest act a human being can sink to”, Silverman has been a vocal backer of Democrat senator Bernie Sanders, Golden’s social media hints at a leaning to the left and the NYT newspaper is left-of-centre.
OpenAI notified on Tuesday that it has purchased a startup “digital product company” Global Illumination.
The New York-based outfit was founded in 2021 by Thomas Dimson, Taylor Gordon and Joey Flynn, who previously designed and built products early on at Instagram and Facebook and have also made significant contributions at YouTube, Google, Pixar and Riot Games.
In its announcement, OpenAI didn’t disclose the terms of the acquisition but said that Global Illumination’s “entire team” has joined the company to work on its “core products,” including ChatGPT.
“Global Illumination is a company that has been leveraging AI to build creative tools, infrastructure, and digital experiences,” OpenAI said in the announcement.
Global Illumination’s sparse website has a brief description of the company, lists eight employees, and has a link to a game called Biomes, which is described as an “open source sandbox MMORPG built for the web” and looks to have been inspired by Minecraft, Roblox and Animal Crossing, looking at the video published on YouTube last month.
The acquisition may be a signal that OpenAI intends to expand its consumer-facing AI products and services, as well as to leverage the expertise and creativity of the Global Illumination team to enhance its existing AI systems.
After OpenAI made US$30 million in revenue last year, chief exec Sam Altman has been reported to have told investors that the company aims for that to surge to around US$200 million this year and US$1 billion in 2024.
Earlier this week, a statement on the London Stock Exchange’s regulatory news service reported that New York investment firm Ripplewood was intending to invest US$1 billion in OpenAI.
However, this turned out to be a bizarre hoax, with the press statement revealed to be fake and Ripplewood’s chief executive Tim Collins calling it “illegal” and saying the firm has raised the issue with authorities.