The latest around generative AI as a whole. Stay up-to-date about current discussions and trends, new findings, and more.
Getty Images has released a high quality image dataset for AI training on Hugging Face. The dataset includes 3,750 images from 15 categories and is characterized by commercial security and high quality, Andrea Gagliano, head of data science and AI/ML at Getty Images, tells VentureBeat. The images are curated specifically for machine learning, in high …
Salesforce is making a radical shift to autonomous AI agents. CEO Marc Benioff announced the new “Agentforce” platform, which will be presented in more detail at the Dreamforce conference. It will allow users to create AI-powered agents that can automate complex customer service tasks. Benioff sees this as the future of customer relationship management and …
The new “Multimodal Arena” from LMSYS compares the performance of different AI models on image-related tasks and shows that OpenAI’s GPT-4o leads the pack, closely followed by Claude 3.5 Sonnet and Gemini 1.5 Pro. Surprisingly, open source models such as LLaVA-v1.6-34B achieve results comparable to some proprietary models. The catch? Despite progress, Princeton’s CharXiv benchmark …
Users can apparently download paid articles from renowned publications such as the New York Times or The Atlantic as HTML files via a feature on Quora’s AI chatbot platform Poe. Experts see this as a potential violation of copyright law. Quora, on the other hand, compares the feature to cloud storage services. The publishers involved …
OpenAI is developing a new way to improve the training of AI models. Until now, human reviewers have been used to check the quality of AI output. Now, an additional AI called CriticGPT will assist the humans.
Amazon is currently investigating whether the Perplexity AI search engine violates the Amazon Web Services Terms of Service. Perplexity is alleged to crawl parts of sites that are blocked via Robots Exclusion Protocol (robots.txt). Read more about the allegations here.
YouTube now allows users to request the removal of AI-generated content that simulates their face or voice. YouTube will review each request on a case-by-case basis, taking into account factors such as the content being flagged as synthetic, the unique identification of the person, and the potential value of the content.
Meta replaces its “Made with AI” tag for photos with “AI Info”. This should eliminate confusion about photos that have only been edited with AI-powered tools.
Google’s DeepMind researchers have developed a new method called JEST that significantly speeds up AI training while reducing energy requirements. By optimizing the selection of training data, JEST can reduce the number of iterations by a factor of 13 and the computational complexity by a factor of 10.
A hacker broke into OpenAI’s internal communications systems last year and stole information about the development of the company’s AI technologies. Employees criticized OpenAI for a lack of security measures and feared the technology could threaten U.S. national security in the future.