Month: February 2024

Sarah Silverman’s lawsuit against OpenAI will advance with some of her legal team’s claims dismissed. The comedian sued OpenAI and Meta in July 2023, claiming they trained their AI models on her books and other work without consent. Bloomberg reported on Tuesday that the unfair competition portion of the lawsuit will proceed. Judge Martínez-Olguín gave the plaintiffs until March 13 to amend the suit.

US District Judge Araceli Martínez-Olguín threw out portions of the complaint from Silverman’s legal team Monday, including negligence, unjust enrichment, DMCA violations and accusations of vicarious infringement. The case’s principal claim remains intact. It alleges OpenAI directly infringed on copyrighted material by training LLMs on millions of books without permission.

OpenAI’s motion to dismiss, filed in August, didn’t tackle the case’s core copyright claims. Although the suit will proceed, the judge suggested the federal Copyright Act may preempt the suit’s remaining claims. “As OpenAI does not raise preemption, the Court does not consider it,” Martínez-Olguín wrote.

The US court system has yet to determine whether training AI large language models on copyrighted work falls under the fair use doctrine. Last month, OpenAI admitted in a court filing that it would be “impossible to train today’s leading AI models without using copyrighted materials.”

The result of Silverman’s OpenAI hearing is similar to one in San Francisco in November when Silverman’s claims against Meta were also slashed down to the core copyright infringement claims. In that session, US District Judge Vince Chhabria described some of the plaintiffs’ dismissed claims as “nonsensical.”

Other groups suing OpenAI for alleged copyright-related violations include The New York Times, a collection of nonfiction authors (a group that grew after the initial lawsuit) and The Author’s Guild. The latter filed its claim alongside authors George R.R. Martin (Game of Thrones) and John Grisham.

This article originally appeared on Engadget at https://www.engadget.com/sarah-silvermans-copyright-infringement-suit-against-openai-will-advance-in-pared-down-form-211456302.html?src=rss

Mozilla is the latest in a long line of tech companies to lay off employees this year. The not-for-profit company is firing around 60 people, which equates to roughly five percent of its workforce. Most of those who are leaving Mozilla worked on the product development team. The news was first reported by Bloomberg. 

“We’re scaling back investment in some product areas in order to focus on areas that we feel have the greatest chance of success,” a Mozilla spokesperson told Engadget in a statement. “To do so, we’ve made the difficult decision to eliminate approximately 60 roles from across the company. We intend to re-prioritize resources towards products like Firefox Mobile, where there’s a significant opportunity to grow and establish a better model for the industry.”

According to an internal memo obtained by TechCrunch, Mozilla plans to pare back investments on several products, including its VPN and a tool that automatically scrubs a user’s personal information from data broker sites. The company announced the latter just one week ago. Hubs, the 3D virtual world Mozilla debuted in 2018, is shutting down while the company is also reducing resources dedicated to its Mastodon instance.

One area into which Mozilla does plan to funnel extra resources is, unsurprisingly, artificial intelligence. “In 2023, generative AI began rapidly shifting the industry landscape. Mozilla seized an opportunity to bring trustworthy AI into Firefox, largely driven by the Fakespot acquisition and the product integration work that followed,” the memo reportedly reads. “Additionally, finding great content is still a critical use case for the internet. Therefore, as part of the changes today, we will be bringing together Pocket, Content and the AI/ML teams supporting content with the Firefox Organization.”

The reorganization comes after Mozilla appointed a new CEO just last week. Former Airbnb, PayPal and eBay executive Laura Chambers, who joined Mozilla’s board three years ago, was appointed chief executive for the rest of this year. “Her focus will be on delivering successful products that advance our mission and building platforms that accelerate momentum,” Mitchell Baker, Mozilla’s former long-time CEO and its new executive chairman, wrote when Chambers took on the job.

This article originally appeared on Engadget at https://www.engadget.com/mozilla-is-laying-off-around-60-workers-210313813.html?src=rss

One of the big drawbacks of talking to an AI chatbot is that everything resets once the conversation is done. It won’t remember who you are or what you previously queried. This is by design, for privacy reasons, but it really hampers the tech from growing into a true digital assistant that knows you well enough to actually help with stuff. 

OpenAI is trying to fix this issue and is finally adding a memory feature to ChatGPT. This will allow the bot to remember important personal details from prior conversations and apply that context to current queries.

Here’s how it works. You can actually tell ChatGPT to remember something specific, like that your child is allergic to peanuts or how you prefer to sign emails. It’ll automatically store this data as requested and apply it to future conversations and tasks.

Beyond that, the system will pick up stuff over time. It’ll naturally store data as you perform queries. The goal is for the chatbot to become smarter and, more importantly, attuned to your specific needs.

Additionally, each custom GPT will have its own unique memory. OpenAI gives Books GPT as an example, as the bot will remember which books you’ve already read and what genres you like. The GPT Store is filled with unique chatbots that would absolutely benefit from this memory feature.

This doesn’t seem too different from how the internet already works. Apps store personal data, as do websites and social networks. This data is then leveraged to create a unique user profile, which is used to develop personalized algorithms. However, there are significant privacy concerns with the standard way of doing things that absolutely carry over to this memory feature.

OpenAI

To that end, OpenAI says users will have control over ChatGPT’s memory and that the system has been trained to not automatically remember certain sensitive topics, like health data. The company says you can simply tell the bot to forget something and it will. There’s also a Manage Memory tab included in the settings for more nuanced adjustments. If the whole idea skeeves you out, just shut the feature off entirely.

This is a beta service, for now, and is rolling out to a “small number” of ChatGPT free and Plus users this week. The company will share plans for a broader release in the future. In the meantime, you can rewatch the movie Her to see where this all ends. 

This article originally appeared on Engadget at https://www.engadget.com/chatgpt-is-getting-a-digital-memory-to-recall-your-past-conversations-195126580.html?src=rss