Authors Win Class Action Approval Against Anthropic Over AI Training Data
A California federal judge has allowed three authors to represent US writers in a class action lawsuit against Anthropic. The claim alleges the AI company trained its models using approximately 7 million books downloaded from pirate libraries such as LibGen and PiLiMi during 2021-2022.
Despite fair use protections for AI training, Anthropic is accused of copyright violations by retaining pirated copies beyond training purposes, creating what the court described as a “central library of all the books in the world.” An Anthropic spokesperson stated the company is considering challenging the ruling, citing the difficulty of proving copyright ownership on such a large scale.
This case adds to mounting legal pressure on AI companies like OpenAI, Microsoft, and Meta regarding the use of copyrighted material in training data.
Meta Rejects EU’s Voluntary AI Code, Citing Legal Uncertainty
Meta has declined to sign the European Union’s new voluntary AI code of practice. Joel Kaplan, Meta’s global affairs chief, argued that the guidelines introduce legal uncertainties and go beyond the scope of the EU’s AI Act.
The voluntary framework is set to take effect next month, but Meta, along with companies like ASML Holding and Airbus, is pushing for a two-year delay in implementation. In contrast, OpenAI recently committed to signing the code.
Kaplan warned that the code could stifle AI innovation and disadvantage European startups compared to larger companies able to handle regulatory complexity.
UK Post Office Warns Horizon IT Scandal Inquiry Could Disrupt £410M System Replacement
The UK Post Office has expressed concerns that the ongoing inquiry into the Horizon IT scandal may require costly changes to its £410 million system replacement project. Instead of building a new system internally, the Post Office opted to purchase software from external suppliers with contracts worth £269 million and £141 million.
Procurement documents highlight that findings from the inquiry “cannot be addressed in the contracts as awarded” and may lead to modifications later. The original Horizon system errors caused wrongful prosecutions of subpostmasters between 1999 and 2015.
This case illustrates how legacy IT failures can create risks that complicate procurement and project planning.
Meta Settles $8 Billion Shareholder Lawsuit Over Cambridge Analytica Privacy Issues
Meta executives have reached a settlement in a multibillion-dollar shareholder lawsuit concerning privacy violations linked to Cambridge Analytica. The settlement, reached before trial testimony resumed, keeps financial terms confidential.
The 2018 lawsuit alleges that Mark Zuckerberg’s decisions allowed Cambridge Analytica to access millions of Facebook users’ data during the 2016 US presidential campaign. Shareholders sought reimbursement from 11 board directors for regulatory fines and legal costs.
High-profile board members involved include Peter Thiel, Reed Hastings, and Jeffrey Zients. The settlement avoids potential exposure of internal discussions and helps protect Meta’s governance reputation amid increasing regulatory scrutiny.
OpenAI and UK Government Partner to Boost AI Infrastructure Investment
OpenAI and the UK government have signed a strategic partnership to enhance AI security research and explore investments in UK data centers and infrastructure. The government plans to invest £1 billion in AI computing, aiming to expand public compute capacity twentyfold over five years.
OpenAI may grow its London office and apply AI technology across sectors including justice, defense, security, and education. CEO Sam Altman praised the UK’s “AI Opportunities Action Plan,” while Technology Secretary Peter Kyle emphasized AI’s role in improving public services and driving economic growth.
The UK government projects AI could increase productivity by 1.5% annually, adding £47 billion of value over the next decade.
Your membership also unlocks:
