AI News

Robocalls are already a nuisance, but the rise of AI-generated voices—like those mimicking President Biden—exacerbates the issue. In response, the FCC is proposing a rule that would make the use of voice cloning technology in robocalls unequivocally illegal, facilitating the prosecution of those behind these fraudulent operations.

In November, Adam D’Angelo became embroiled in one of the most significant controversies in the tech sector. The OpenAI board — overseeing a $80 billion startup at the forefront of artificial intelligence — unexpectedly dismissed its CEO, Sam Altman, only to reinstate him days later. D’Angelo was among the board members who voted for Altman’s removal and is still part of the board that reinstated him. Remarkably, he was the only original member to retain his position during the board's subsequent restructuring.

India has recently revised its AI advisory following backlash from local and global entrepreneurs and investors. The Ministry of Electronics and IT presented an updated AI advisory to industry stakeholders on Friday, removing the previous requirement for government approval before launching or deploying AI models in the South Asian market.

Last month, Amazon unveiled its new AI-powered chatbot, Rufus, integrated into the Amazon Shopping app for both Android and iOS. Following a brief delay, the rollout for early testers commenced on February 1. Rufus aims to enhance the shopping experience by helping users find, compare, and receive tailored product recommendations. Naturally, I had to put it to the test.

Navigating the Pitfalls of Generative AI: A New Approach to Risk Management

Clearview AI Wins Appeal Against UK Privacy Sanction

A new automation startup is ready to disrupt the market, officially launching to the public today after an extended beta period. Relay, as the company is named, positions itself as a comprehensive workflow automation platform that assists individuals in managing tedious, repetitive tasks, going “beyond triggers and actions” commonly associated with established platforms like Zapier and IFTTT.

At Google’s I/O developer conference, the tech giant presented its vision for the future of AI, aiming to establish its superiority over competitors. Attendees witnessed the launch of a new AI-enhanced search engine, an AI model that boasts an expanded context window of 2 million tokens, and AI-driven tools integrated into its Workspace apps like Gmail, Drive, and Docs. Furthermore, Google introduced tools for developers to incorporate its AI into their applications and offered a glimpse into Project Astra, an ambitious AI project designed to interact using sight, sound, voice, and text.

Elon Musk’s AI Startup xAI to Open Source Grok Chatbot This Week

In a dramatic weekend at OpenAI, investors and some employees made a last-ditch effort to persuade the board to reinstate former Y Combinator president Sam Altman after his sudden dismissal on Friday. According to a report by The Information, Altman will not be returning as CEO. An internal memo from Ilya Sutskever, a board director and co-founder of OpenAI, confirmed that Altman has decided to step away from negotiations for the time being. As OpenAI searches for a new permanent CEO, Emmett Shear, the co-founder of Twitch, has been appointed as the interim CEO, taking over from Mira Murati, who served just two days in the role.

AI systems and large language models (LLMs) require extensive datasets for accuracy, but it’s crucial that they’re trained only on data they have the rights to use. Recent licensing agreements between OpenAI and media outlets like The Atlantic and Vox indicate a growing interest from both parties in establishing content licensing deals for AI training.

On Thursday, Reddit announced the launch of a new policy designed to balance its goals of licensing content to major tech companies, such as Google, while safeguarding user privacy. This newly introduced “Public Content Policy” will complement Reddit’s current privacy and content policies, detailing how third parties can access and utilize Reddit’s data. Additionally, the company revealed the creation of a dedicated subreddit for researchers interested in working with Reddit’s data.

The highlight of YouTube’s Made On YouTube event on Wednesday morning was undoubtedly artificial intelligence. The platform unveiled its integration of Google DeepMind’s AI video generation model, Veo, into YouTube Shorts, empowering creators to produce high-quality backgrounds and six-second clips effortlessly.

Japan is forging new partnerships with U.S. universities to accelerate advancements in artificial intelligence (AI) research. Under the collaborative agreements, the University of Washington and the University of Tsukuba will work together on AI initiatives, specifically focusing on workplace development. Concurrently, Carnegie Mellon University will partner with Keio University to explore multimodal learning, embodied AI, and applications in life sciences.

Convolutional Neural Networks (CNNs) have long been recognized as the leading architecture for tasks in computer vision, particularly for image classification. Recently, Vision Transformers (ViTs) have emerged as a compelling alternative due to their enhanced performance in accuracy and efficiency when scaled. However, research from Google DeepMind reveals that both CNNs and ViTs can achieve comparable results, with the amount of computing power employed during training being the critical factor.