AI Training

On January 6th, according to The Inside of Business, the data label company Invisible Technologies CEO Matt Fitzpatrick stated in the podcast 20VC that a core misunderstanding surrounding the generation of AI training was gradually being corrected: the role of humans in data production and feedback would not be replaced by synthetic data in the coming decades. Fitzpatrick points out that there is a long-standing view within the industry that AI..

On 30 October, according to information received, this week Meta submitted an application to the United States District Court to dismiss a lawsuit alleging that it had illegally downloaded pornographic material for use in AI training through seed. According to TorrentFreak, Strike 3 Holdings had previously discovered that Meta's company IP address contained illegal downloads of some of its adult films, while it claimed that Meta had covered up other downloads through a “secret network” of 2,500 “hidden IP addresses”. Strike 3..

April 24, 2011 - Adobe has launched Content Authenticity, a new app that protects creative workers' rights to attribute their work, even if someone takes a screenshot and redistributes it across the web. The web app, which is available for public beta testing today, allows users to upload images that need to be protected, and the app embeds invisible metadata and adds a "No AI Training" tag to help confirm attribution and protect the work from AI training. The new web app is based on Adobe's Content Credential...

According to NVIDIA's official blog on January 13, NVIDIA announced the launch of a large English-language AI training database called Nemotron-CC, which contains a total of 6.3 trillion Token, of which 1.9 trillion are synthetic data. NVIDIA claims that the training database can help further the process of training large language models for both academia and the corporate world. Currently, the specific performance of various AI models in the industry depends largely on the training data of the respective models. However, existing publicly available databases are often limited in size and quality, and NVIDIA claims that Nemotron-...

Recently, social platform X updated its privacy policy, which will allow X platform to use user data to train AI models from November 15, unless the user opts out, triggering user dissatisfaction. Previously, Adobe, Google and other companies also introduced similar content in the terms and conditions, causing controversy over the conflict between AI training and privacy, copyright, etc., and related legal issues are still under discussion. Change: user data will be used for AI training Recently, the X platform updated its privacy policy with a new clause that allows it to share user data with third parties to train AI, unless the user opts out. However, the platform did not provide a clear opt-out option and reminded users that even within...

"Retarded Bar" is a sub-forum of Baidu Post Bar. In this forum, users create content that contains puns, multiple meanings, inversions of cause and effect, and harmonic words that are either brilliant or "brain-teasing". Some of the content has logical traps that are challenging even for humans. In April of this year, a team of researchers from the Chinese Academy of Sciences, in a study titled "COIG-CQIA: Quality is All You Need for Chinese Instruction Fine-tuning," used the phrase "retarded title + GPT-4 answer" to describe a "GPT-4 answer" to a question. GPT-4 answer "Fine-tuning...

Adobe has drawn strong criticism from users for updating its terms of service in the past few weeks. On Tuesday, the company announced revised terms of service that explicitly emphasize that it will not use user content stored locally or in the cloud to train AI models. The updated terms add several new subcategories to the section on Adobe's access to user content, one of which is specifically for generative AI. The terms clearly state that Adobe's software "will not use your local or cloud content to train generative AI models." However, there is one exception: if you submit your work to Adobe Sto…

In a recent interview with Axios, U.S. Energy Secretary Jennifer Granholm said that she is actively exploring nuclear power plant solutions to alleviate the electricity demand of AI data centers. Granholm said that the development of AI in the United States is not a problem in itself, and AI can help humans solve many problems. But the key to the problem is that the rapid development of AI has exceeded the carrying capacity of the national power grid, and the existing power grid cannot support the rapid expansion of AI data centers. According to reports, the U.S. Department of Energy (DOE) has been working with Microsoft, Google, and...