WIRED | Swipe Insight

Meta Faces Legal Setback Over AI Training with Pirated Books from Library Genesis

1 months ago

Meta is in legal trouble after court documents revealed it trained its AI models using pirated books from Library Genesis, a notorious piracy site. Authors Richard Kadrey and Sarah Silverman filed a lawsuit alleging copyright infringement. A judge criticized Meta's redaction efforts, indicating they were to avoid negative publicity. Meta claims its use of publicly available materials is protected under "fair use." The case's outcome could affect future AI training and copyright laws.

Top-Notch Google Ads Audit Tool

Featured

Effortlessly audit your Google Ads account with Marketing Auditor. Perform 200+ automated checks to uncover optimization opportunities and save over 10 hours per audit. Generate white-label reports in minutes with 50+ pages of actionable insights. Customize your reports with professional themes or your own branding, and export them in editable formats like PowerPoint or Google Slides. This tool is the ultimate solution for efficient and impactful Google Ads audits.

Perplexity AI Accused of Plagiarism and Misleading Practices

8 months ago

Perplexity

A WIRED investigation reveals Perplexity, an AI search startup funded by Jeff Bezos' family fund, Nvidia, and Balaji Srinivasan, is accused of stealing content and fabricating information. Despite claims of reliable answers, Perplexity allegedly scrapes websites without permission and misrepresents sources. Forbes accused it of plagiarism, leading to scrutiny. Perplexity's CEO denies wrongdoing but did not dispute specifics.

How to Use ChatGPT’s Memory Feature

9 months ago

ChatGPT

OpenAI has recently introduced a new feature called "Memory" for its ChatGPT Plus. This feature allows the program to store personal information shared during conversations for future reference. However, it's currently only available to Enterprise and Team customers, and not yet integrated with the GPT Store. While it has been tested by a small group, it's not currently available in Europe or Korea.

Trending ️‍🔥

A Breakthrough Online Privacy Proposal Hits Congress

10 months ago

Data Protection

The American Privacy Rights Act (APRA), a new bipartisan proposal in Congress, aims to limit the types of consumer data companies can collect, retain, and use. It proposes creating a national registry of data brokers and requires these companies to let users opt out of having their data sold. The APRA includes provisions from California’s privacy law and exempts small businesses, governments, and certain nonprofits. However, its approval remains uncertain.

Meta Kills a Crucial Transparency Tool At the Worst Possible Time

11 months ago

Apple’s MM1 AI Model Shows a Sleeping Giant Is Waking Up

11 months ago

AI Apple

Apple, a tech giant that has been slow to adopt generative AI, is reportedly in early discussions with Google about incorporating Google's Gemini AI model into iPhones. This information comes from a recent New York Times report and earlier Bloomberg coverage. Despite the tech industry's enthusiasm for AI, Apple has not yet introduced any AI-generated features, such as emojis.

Reddit’s Sale of User Data for AI Training Draws FTC Inquiry

11 months ago

Reddit Organic Social

Reddit's plan to license user posts for AI projects, potentially earning $203 million, has drawn FTC scrutiny. The FTC is investigating the sharing of user-generated content with third parties for AI training. Reddit insists it hasn't engaged in unfair practices. The FTC's scrutiny is part of a non-public inquiry. Other companies like Stack Overflow and Tumblr are also licensing user data for AI.