Search AI Products and News

Joelle Pineau, VP of Meta AI Research, announced on Tuesday via a Facebook post that she will be leaving her position at Meta in May. Pineau has served as the head of Meta's Fundamental AI Research (FAIR) lab for the past two years, leading the lab's cutting-edge research in artificial intelligence. FAIR is a core internal research team at Meta, led by renowned scientist Yann LeCun. Pineau's departure comes as Meta...

Meta is testing a new feature that allows users to generate comment suggestions on Instagram using artificial intelligence. X user Jonah Manzano discovered and shared this test feature, showing users can tap a pencil icon below a post to access Meta AI, which analyzes the photo and provides three possible comments. Users can refresh for more suggestions if unsatisfied. A Meta spokesperson confirmed the company is testing Meta AI in multiple areas.

After a nearly year-long pause, Meta has announced the return of its AI chatbot, Meta AI, to the European market. Starting this week, Meta AI will be rolled out across WhatsApp, Facebook, Instagram, and Messenger in 41 European countries and 21 other regions, but initially only with text-based chat functionality. Meta AI initially launched in the US in 2023. Although Meta had planned to bring the assistant to Europe, it was delayed due to concerns in Ireland.

Social media giant Meta is facing unprecedented AI infrastructure costs, with projected AI-related spending soaring to a staggering $65 billion, potentially contributing to a total annual expenditure of $119 billion! Faced with this astronomical bill, the tech giant has decided to develop its own AI chips, and has already made significant progress. Recent reports indicate that Meta is about to begin a small-scale deployment of its custom chips, signaling a gradual move away from Nvidia and its reliance on their GPUs.

CNBC reports that Meta is planning a standalone AI assistant app called Meta AI to better compete with OpenAI's ChatGPT and Google's Gemini. The app is expected to launch in the company's next fiscal quarter, between April and June 2025. Currently, Meta AI is only accessible through the company website and Meta's social media apps like Facebook and WhatsApp.

Recently, the Meta AI team launched the video joint embedding prediction architecture (V-JEPA) model, an innovative initiative aimed at advancing machine intelligence. Humans can naturally process information from visual signals and recognize surrounding objects and motion patterns. An important goal of machine learning is to reveal the fundamental principles that drive unsupervised learning in humans. Researchers proposed a key hypothesis—the predictive feature principle—arguing that the representations of continuous sensory inputs should be able to predict each other. Early research methods utilized slow feature analysis.

Recently, Meta CEO Mark Zuckerberg's alliance with Trump and his significant denial of European values have raised a troubling question: Should European organizations continue to use Meta's AI models? Not long ago, Meta criticized the EU for blocking its use of European user data to train AI models, claiming this data is crucial for localizing technology. However, Zuckerberg's recent statements are at odds with these values, as he announced plans to cooperate with the Trump administration to oppose so-called external governance.

Yann LeCun, head of AI research at Meta, recently stated in an interview that current AI systems are not as powerful as some researchers have claimed. He noted that the true understanding, planning, and reasoning capabilities of AI are still lacking and that multiple 'conceptual breakthroughs' are needed to achieve human-like intelligence. Image source note: Image generated by AI, image licensed from provider Midjourney. LeCun criticized the overly optimistic attitudes of OpenAI and Google DeepMind towards AI.

In recent years, large language models (LLMs) have made significant progress in the field of natural language processing (NLP), widely applicable in scenarios such as text generation, summarization, and question answering. However, these models rely on a token-level processing method that predicts word by word, which struggles with contextual understanding and often leads to inconsistent outputs. Moreover, when scaling LLMs to multilingual and multimodal applications, the computational costs and data requirements tend to be relatively high. To address these issues, Meta AI has proposed a novel approach.