AI news & updates

A recent study claims OpenAI trained its AI models using paywalled O'Reilly books without permission, raising significant copyright concerns. This accusation adds to ongoing debates about the ethical use of copyrighted materials in AI training.

Joelle Pineau, Meta's VP of AI research, is set to leave the company in May after over two years leading the FAIR lab. Her departure coincides with Meta's intensified focus on AI advancements.

OpenAI CEO Sam Altman announced that the surge in demand for ChatGPT's new image-generation tool will lead to delays in product releases, indicating ongoing capacity challenges within the company.

NaNoWriMo, a nonprofit writing community, is shutting down due to scandals involving AI and content moderation, raising concerns about the impact of AI on creative platforms and community integrity.

Qualcomm has acquired the generative AI division of VinAI, a Vietnamese AI research firm founded by a former DeepMind scientist. This acquisition signifies Qualcomm's strategic move to enhance its AI capabilities.

Tinder has launched an AI-powered game that allows users to practice flirting with an AI bot, aiming to enhance dating skills through interactive scenarios and personalized feedback.

While OpenAI's ChatGPT remains the leading chatbot, competitors like Google's Gemini and Microsoft's Copilot are experiencing significant user growth, as indicated by traffic data from Similarweb and Sensor Tower.

Brain-computer interfaces (BCIs) are advancing, enabling paralyzed individuals to control computers through imagined movements. Additionally, a clinical trial shows generative AI therapy bots can benefit those with mental health issues, emphasizing the importance of training data.

Brain-computer interfaces (BCIs) are advancing with clinical trials from companies like Neuralink and Synchron, aiming to enable paralyzed individuals to control computers. Despite progress, practical deployment remains a challenge, with questions about longevity and usability of implants.

Researchers at Dartmouth College developed Therabot, an AI therapy model, emphasizing the importance of training data. Initial attempts using online conversations failed, but using therapy session transcripts and cognitive behavioral techniques improved outcomes, highlighting the need for evidence-based training in AI therapy.

A group of 50 sports fans utilized hyperchat technology to collaboratively create a March Madness bracket, achieving remarkable accuracy by performing in the 99th percentile of a major contest, showcasing the potential of AI-enhanced group intelligence.

Emergence AI has launched a real-time, no-code platform that autonomously creates AI agents based on user-defined tasks. This innovative system enhances enterprise automation by integrating large language models with autonomous agent technology.

The iPhone 16 Pro introduces Apple Intelligence features like Notification Summaries and Camera Control, enhancing user experience. However, the AI capabilities are still seen as underwhelming compared to competitors.

OpenAI has made its image generation feature available to free ChatGPT users, allowing them to create images directly within the platform. Users can generate up to three images daily, with popular requests including Ghibli-style transformations of photos, raising copyright concerns.

UK authors, led by the Society of Authors, demand accountability from Meta for alleged copyright infringement involving its AI model, Llama 3, which reportedly used pirated materials for training. They urge the government to protect authors' rights against corporate exploitation.