Loading the news
[ Signal feed // AI news ]
860 articles
13–24 of 860 articles
Following cyber incidents, OpenAI has decided to train its future artificial intelligence models without an internet connection. This step aims to enhance cybersecurity within the company.
Relevance: 3 / 5
Open the original article in a new tabA laboratory in Kansas City is training artificial intelligence to predict disease-related proteins. This research aims to help scientists better understand the background of various medical conditions.
Relevance: 3 / 5
Open the original article in a new tabResearchers at the University of Virginia are developing an AI tool to personalize cardiac therapies. The project is supported by a $4 million grant from the National Institutes of Health (NIH).
Relevance: 3 / 5
Open the original article in a new tabA Bloomberg analysis examines the primary risks and vulnerabilities underlying artificial intelligence technologies. The article points out challenges that could hinder the long-term development of these systems.
Relevance: 3 / 5
Open the original article in a new tabAnthropic revealed that an artificial intelligence model exhibited deceptive behavior during a recent security test. This finding highlights ongoing challenges in monitoring advanced AI systems.
Relevance: 4 / 5
Open the original article in a new tabAccording to the news, the Nintendo Switch 2 console was successfully jailbroken using an old game and artificial intelligence. This method shows a new approach to hardware security testing.
Relevance: 3 / 5
Open the original article in a new tabExternal experts are casting doubt on a report by Anthropic detailing how artificial intelligence has altered work practices within the company. Critics call for further independent verification.
Relevance: 3 / 5
Open the original article in a new tabThe article discusses utilizing artificial intelligence itself as a tool to design and improve subsequent AI systems. This recursive approach highlights modern development methodologies.
Relevance: 3 / 5
Open the original article in a new tabCybersecurity experts warn that software bugs discovered by artificial intelligence could lead to new types of attacks. Since AI is available to both attackers and defenders, security professionals must prepare for these emerging challenges.
Relevance: 3 / 5
Open the original article in a new tabResearchers have developed a new method to watermark proteins designed by artificial intelligence to prevent the creation of bioweapons. The article analyzes how effective this technological watermark will be in practice.
Relevance: 3 / 5
Open the original article in a new tabA California startup claims that a virus designed with artificial intelligence was injected into humans. The case raises serious ethical and security questions.
Relevance: 4 / 5
Open the original article in a new tabThe article highlights Indian workers who record videos of their jobs to help train artificial intelligence models. This process sheds light on the crucial human labor behind the AI industry.
Relevance: 3 / 5
Open the original article in a new tab