Here’s what you’ll find in this fortnight’s AI updates:
- The New York Times files a lawsuit against Microsoft and OpenAI for alleged copyright infringement
- Chief Justice John Roberts publishes Year End Report on AI and Judiciary
- Bill seeking to regulate foundation models introduced in the US Congress
- Jio to soon launch BharatGPT program in collaboration with IIT Bombay
- Music creation feature introduced in Microsoft Copilot via Suno Integration
- Google Cloud announces text to image generator, Imagen 2
THE NEW YORK TIMES FILES A LAWSUIT AGAINST MICROSOFT AND OPENAI FOR ALLEGED COPYRIGHT INFRINGEMENT
OpenAI is entangled in a legal dispute with the New York Times along with Microsoft over alleged copyright infringement in the development of AI training models. Filed on December 27, 2023, in the United States District Court Southern District Of New York, The Times claims that Microsoft’s Copilot and OpenAI’s ChatGPT utilized its copyrighted content without authorization, jeopardizing its services. Despite failed negotiations for fair compensation and responsible GenAI development, the defendants argue “fair use,” while The Times contends it as unfair competition. The lawsuit seeks statutory and compensatory damages, along with an injunction against further unlawful conduct, highlighting a clash between ethical journalism and AI advancement.
CHIEF JUSTICE JOHN ROBERTS PUBLISHES YEAR END REPORT ON AI AND JUDICIARY
In the 2023 Year-End Report on the Federal Judiciary, Chief Justice John Roberts discussed the transformative impact of artificial intelligence (AI) on the legal system. While recognizing AI’s potential, he affirmed the enduring relevance of human judges. Roberts expressed confidence in the coexistence of human judges and AI, particularly at the trial level. He noted the rise of AI capable of passing the bar exam but cautioned against generating fictitious content. Emphasizing the need to adapt to this technological shift, the Chief Justice outlined challenges and opportunities, addressing privacy concerns and the risk of dehumanizing the law. As AI evolves, federal courts must carefully assess its applications, involving various committees in decision-making.
BILL SEEKING TO REGULATE FOUNDATION MODELS INTRODUCED IN US CONGRESS
On December 22, US representatives Anna Eshoo and Don Beyer, chairs of the Congressional AI Caucus, introduced the AI Foundation Model Transparency Act. The Bill targets “covered entities,” defined as those generating over 100,000 monthly instances or having over 30,000 monthly users using foundation models (FMs). It urges the FTC to establish transparency standards mandating covered entities to disclose FM training data, mechanisms, and user data practices. Violations are deemed “unfair or deceptive acts” prosecutable by the FTC. Critics highlight vague terms and a limited scope, potentially favouring big tech. A more flexible approach is suggested for broader impact and effective regulation.
JIO TO SOON LAUNCH BHARATGPT PROGRAM IN COLLABORATION WITH IIT BOMBAY
Reliance Jio announced its collaboration with IIT Bombay for the BharatGPT program, a pivotal project. At IIT Bombay’s Techfest, Chairman Akash Ambani revealed AI’s imminent widespread integration across sectors like commerce, media, education, healthcare, and agriculture. Jio intends to embed AI not only vertically but horizontally across operations, focusing on telecom and retail AI solutions. The BharatGPT program, a forthcoming AI system, aims to revolutionize Jio’s telecom products through voice and gesture activation.
MUSIC CREATION FEATURE INTRODUCED IN MICROSOFT COPILOT VIA SUNO INTEGRATION
Microsoft introduced a new plugin feature on its Copilot platform, called Suno, which enables users to generate complete songs from prompts. The Suno plugin generates songs from start to finish complete with lyrics, instrumentals, and vocals, based on text prompts the user enters into it. Users simply have to write prompts and the AI generates music based on the prompt. The tool requires no musical prowess or technical ability to be able to create music. As of now, Suno has not revealed its sources of AI training data nor does it restrict users from certain prompts including those that allow the user to generate content in the voice of certain musicians.
The trend of AI-generated music is gaining momentum with Suno and the YouTube feature allowing users to generate music. In late 2023, the Grammys announced their ban on generated music content, while the RIAA submitted a proposal to the US Trade Representative asking them to declare AI-generated music as pirated content.
GOOGLE CLOUD ANNOUNCES TEXT TO IMAGE GENERATOR, IMAGEN 2
Google Cloud has launched Imagen 2. It is an upgrade to its text-to-image capabilities for Vertex AI customers. Developed with Google DeepMind tech, Imagen 2 offers enhanced image quality and tailored functionality. Key features include diverse image generation, text rendering in multiple languages, logo creation, and advanced captions or Q&A. It supports 6 additional languages in preview and incorporates safety measures aligned with Google’s Responsible AI principles. Imagen 2, embraced by companies like Snap, Shutterstock, and Canva, and is designed to meet enterprise standards, providing reliability and governance. Google Cloud expects organizations to leverage Imagen 2 for creative image generation.
Authors: Astha Singh, Shruti Gupta
