Twice a week, Data Points brings you the latest AI news, tools, models, and research in brief. In today’s edition, you’ll find:
- U.S. shuts down more chips and tech to China
- Claude’s Google Docs integration
- Adobe’s MultiFoley generates sound for video
- Canadian media companies sue OpenAI
But first:
OpenAI explores advertising for its AI products amid revenue push
OpenAI’s CFO Sarah Friar revealed the company is considering an advertising model for its AI products, though it has no immediate plans to implement ads. The $150 billion-valued startup has been hiring advertising experts from Meta and Google, including Shivakumar Venkataraman, former leader of Google’s search advertising team. OpenAI’s revenue has surged to about $4 billion annually, largely due to ChatGPT’s success, which now has over 250 million weekly active users. However, the company anticipates spending more than it earns in the near term, with cash burn expected to exceed $5 billion, as it continues developing advanced AI models. (Financial Times)
Cohere releases new enterprise search model Rerank 3.5
Cohere introduced Rerank 3.5, an AI model designed to improve information retrieval in search and retrieval-augmented generation systems. The model aims to enhance reasoning capabilities, handle various data types, and perform better across multiple languages. Rerank 3.5 uses a cross-encoding method to calculate relevance scores between user questions and documents. The release may interest businesses looking to refine their AI-powered search systems, particularly in specialized industries like finance and healthcare. (Cohere)
U.S. tightens restrictions on advanced chip exports to China
The Biden administration announced new restrictions on technology exports to China, focusing on advanced chips and semiconductor manufacturing equipment. The rules ban sales of certain AI chips, advanced memory chips, and 24 types of semiconductor equipment to China. Additionally, 140 Chinese companies, many involved in chip-making tools and machinery, were added to a restricted trade list. The restrictions also cover specific software tools used in chip development and will apply globally to prevent offshore workarounds. These measures aim to impede China’s ability to produce cutting-edge chips for military and AI applications. (The New York Times)
Claude gains ability to read Google Docs
Anthropic added Google Docs integration to Claude, allowing users to connect documents directly to conversations and projects. The feature extracts text from Google Docs, enabling Claude to access up-to-date document content for improved context and assistance. This integration enhances Claude’s ability to understand and assist with complex tasks by incorporating relevant information from users’ Google Drive documents. (Anthropic)
New model generates custom sound effects for videos
Adobe researchers introduced MultiFoley, an AI model that creates sound effects for videos using text, audio, and video inputs. The system can produce various sounds, from realistic to imaginative, and allows users to reference existing audio or partial videos. Researchers evaluated MultiFoley through automated tests and human studies, comparing its output to existing methods in terms of synchronization and overall audio quality. The results indicate that MultiFoley outperformed other approaches in generating synchronized, high-quality sounds across various input conditions. (arXiv)
Canadian news companies sue OpenAI for copyright infringement
Five major Canadian news organizations filed a lawsuit against OpenAI, accusing the company of using their content without permission or compensation to train its AI systems. The media companies, including Torstar and CBC/Radio-Canada, are seeking damages and a permanent injunction to prevent OpenAI from using their material without consent. This case joins a growing number of lawsuits against AI companies by content creators and publishers, highlighting the ongoing debate over fair use of copyrighted material in AI training. (Reuters)
Still want to know more about what matters in AI right now?
Read last week’s issue of The Batch for in-depth analysis of news and research.
Last week, Andrew Ng shared his gratitude for Thanksgiving, reflected on the struggles of those less fortunate, and emphasized the importance of understanding diverse perspectives to create impactful technology. He highlighted his optimism about AI’s potential to improve lives and encouraged the community to continue building solutions to help others.
“To make good decisions, I have to understand the people I hope to serve. This is why I continue to routinely seek out, speak with, and try to understand people from all walks of life, and I hope many others in AI will do so, too.”
Read Andrew’s full letter here.
Other top AI news and research stories we covered in depth: DeepSeek-R1 challenges OpenAI o1 with a transparent model revealing its reasoning; π0 advances household robotics with an innovative machine learning system; Amazon deepens its partnership with Anthropic through a $4 billion investment; and Grounding DINO 1.5 enhances object detection on small devices with faster and smarter capabilities.