The latest around generative AI as a whole. Stay up-to-date about current discussions and trends, new findings, and more.
YouTube is introducing a new label to identify authentic videos. According to a report by Umar Shakir in The Verge, the “captured with a camera” label will mark videos recorded with a real camera and unaltered footage and sound. This requires the recordings to meet the C2PA standard, version 2.1 or higher. The new feature …
Anthropic has updated its AI safety policy to prevent misuse, reports VentureBeat author Michael Nuñez. The new “Capability Thresholds” define benchmarks for risky capabilities of AI models, such as in the area of bioweapons or autonomous AI research. If a model reaches such a threshold, additional safeguards are triggered. The revised policy also sets out …
The New York Times has demanded that the AI startup Perplexity cease using its content. According to a Wall Street Journal report the newspaper accuses Perplexity of copyright infringement. Perplexity uses the Times’ content for AI-generated search results. The Times is demanding compensation and the surrender of illegal profits. Perplexity CEO Aravind Srinivas said they …
U.S. defense and security agencies have spent $700 million on AI projects since ChatGPT’s launch. According to a Fortune analysis by Kali Hays, the Department of Defense (DOD) and the Department of Homeland Security (DHS) have contracted hundreds of companies to develop and test AI algorithms and tools. The DOD awarded roughly $670 million to …
Microsoft’s Vice President of GenAI research, Sebastien Bubeck, is leaving the company to join OpenAI, according to Reuters. Bubeck will continue his work on developing Artificial General Intelligence (AGI) at the Microsoft-backed AI startup. His specific role at OpenAI remains unclear. Reuters also notes that most of Bubeck’s co-authors on a research paper about Microsoft’s …
Google will back the construction of seven small nuclear reactors in the U.S. to power its growing AI operations, according to Jennifer Hiller in the Wall Street Journal. The tech giant has agreed to purchase the power generated by the reactors, which will be built by nuclear startup Kairos Power. The deal aims to add …
Decentralized training of a 10-billion-parameter model called INTELLECT-1 has begun. Anyone can contribute computing power and participate. INTELLECT-1 is based on the Llama-3 architecture and is trained on a high quality open source dataset called Fineweb-Edu by Hugging Face. The dataset contains over six trillion tokens and consists of Fineweb-edu (55%), DLCM (20%), Stack v2 …
Garry Tan, President & CEO of Y Combinator, addresses recent criticisms directed at the start-up PearAI regarding its handling of open source practices, particularly in relation to the company Continue. Initially, Tan supported PearAI without fully understanding the situation, but he later recognized the validity of the criticism, he writes. He collaborated with PearAI’s founders …
OpenAI’s recent report highlights the misuse of its platform by cyber actors aiming to influence global elections. The company has disrupted over 20 deceptive operations that utilized AI-generated content, including fake articles and social media posts, writes Hayden Field for CNBC. With significant elections approaching worldwide, including the U.S. presidential election, concerns about election-related misinformation …
Walmart is developing its own large language model called Wallaby, tailored specifically for retail applications. Trained on extensive data from Walmart, Wallaby aims to enhance customer service by responding in a natural tone, Emilia David reports for VentureBeat. Currently in heavy internal testing, Walmart plans to leverage Wallaby within a year, particularly for consumer-facing applications. …