DeepSeek, in partnership with Tsinghua University, unveils DeepSeek-GRM, an AI framework that improves reasoning through self-principled critique tuning and generative reward modeling, aiming for more efficient and intelligent AI systems.
A joint study from leading universities reveals OpenAI’s GPT-4 memorized copyrighted fiction and news content, escalating legal tensions over AI training data.
Midjourney V7 introduces voice prompting, faster Draft Mode, improved realism, and a new architecture, while legacy features remain missing and user reactions vary.
Microsoft enhances Copilot with memory, personalization, web-based actions, and real-time visual understanding—marking a leap toward deeply personal AI.
Amazon projects Rufus will add $700 million to its operating profits in 2025 through product recommendations and embedded ads, despite mixed reviews.
CURIE, SPIQA, and FEABench introduce rigorous benchmarks to evaluate LLMs on complex scientific tasks like long-form reasoning, multimodal analysis, and simulation.
Papa Johns and Google Cloud announce a partnership to utilize AI technologies, aiming to enhance customer personalization and streamline delivery processes.
Artificial Intelligence is set to become a $4.8 trillion global market by 2033. UNCTAD’s latest report warns that without immediate measures, the benefits may be confined to a privileged few, exacerbating global inequalities.
UK authors protested Meta’s alleged use of over 7.5 million pirated books from LibGen for AI training. The protest gained momentum across social media and in courtrooms, raising ethical and legal questions about AI data sourcing.
