AI News December 2025: Monthly Digest

Scientific American explores whether today's AI models, which can already write and refine their own software, represent the first steps toward true superintelligence. The piece examines whether this self-improvement capability could eventually snowball into something far beyond human intelligence.

Anthropic released Claude Opus 4.5, which reportedly beat all human job candidates in the company's internal engineering tests, setting new AI capability records. The model comes with improved coding abilities and expanded features like Claude for Excel and Chrome integration. This marks Anthropic's third major model launch in just two months.

Manus Slides has integrated Google's Nano Banana Pro model to generate visually stunning, presentation-ready slide decks from simple prompts, offering studio-quality visuals, flawless text rendering, and cohesive artistic direction. This integration accelerates workflow, democratizes professional design, and maximizes the impact of presentations, enabling users to create high-end visual communication effortlessly and make their message more memorable.

OpenAI CEO Sam Altman sent an internal memo declaring 'code red' as ChatGPT's user growth has slowed significantly. The company is scrambling to improve personalization, reliability, and image generation to compete with Google's Gemini 3 launch.

Databricks raised $10 billion at a $62 billion valuation in December 2024, marking the largest venture capital raise of the year and one of the largest on record. This represents a 44% increase from its 2023 valuation of $43 billion. The funding includes significant investments from Thrive Capital, Andreessen Horowitz, and new investors like Abu Dhabi's MGX.

OpenAI released GPT-5.1 as an upgrade to GPT-5, introducing new personality options like Professional, Candid, and Quirky alongside the existing Default, Friendly, and Efficient modes. The update includes two distinct processing modes - Instant and Thinking - that are automatically selected based on user intent. The model shows improvements in speed, reasoning depth, and task reliability.

The 'Godfather of AI' Geoffrey Hinton agrees with predictions from Bill Gates and Elon Musk that AI will cause widespread job displacement. Hinton is increasingly vocal about Big Tech's priorities and the potential for most jobs to be automated away.

Google introduced Gemini 2.0, its most capable model yet, designed for the new agentic era. With new advances in multimodality including native image and audio output and native tool use, it will enable building AI agents that bring them closer to a universal assistant. The model is now available to all Gemini users experimentally.

Chinese internet companies are significantly increasing their AI investments, potentially benefiting local technology players. This spending surge reflects the competitive landscape in China's AI market and the push for technological self-reliance.

Anthropic acquired Bun, a fast JavaScript runtime and toolkit, to enhance its coding infrastructure as Claude Code reaches $1 billion in usage milestone. Bun will remain open source and MIT-licensed while helping build next-generation software infrastructure. The acquisition positions Anthropic as a major player in developer tools.

OpenAI developed a new version of Sora—Sora Turbo—that is significantly faster than the model previewed in February. They're releasing it as a standalone product at Sora.com to ChatGPT Plus and Pro users. Users can generate videos up to 1080p resolution, up to 20 seconds long, and in widescreen, vertical or square aspect ratios.

Amazon released its new AI Trainium3 chip, claiming it's four times faster and more energy efficient than previous generations. This is part of the broader industry push to break NVIDIA's stranglehold on AI computing hardware.

A new study found that AI chatbots using inaccurate information were able to influence and change people's political opinions. The research highlights concerns about AI's potential impact on democratic processes and information integrity.

Anthropic CEO Dario Amodei criticized OpenAI and Google while his company reportedly prepares for a massive IPO that could rival OpenAI's own public offering plans. The company is expanding beyond enterprise coding into finance, biomedical, retail, and energy sectors. Investment banks have had preliminary discussions about the potential listing.

Anthropic is reportedly in early talks for one of the largest IPOs as early as next year, engaging Wilson Sonsini Goodrich & Rosati law firm. This sets up a potential public market race with OpenAI for AI dominance.

Google's Willow quantum chip achieved a historic milestone by solving a computational problem in under five minutes that would take the world's fastest supercomputers a staggering 10 septillion years. To put this in perspective, 10 septillion years vastly exceeds the age of the universe. This demonstrates exponential quantum supremacy and earned recognition as a 'Physics Breakthrough of the Year.'

Researchers developed an AI model that can predict a person's risk of developing breast cancer up to five years in advance by analyzing mammograms. The breakthrough could revolutionize breast cancer screening by making it much more effective and targeted. The AI algorithm shows high accuracy in risk classification.

Many customers aren't getting enough from Copilot to justify the cost. 'Am I getting $30 of value per user per month out of it?' The short answer is no, and that's what's been holding further adoption back. Microsoft is responding by launching a new Business tier at $21 per person per month for organizations with up to 300 users.

During internal safety testing, Anthropic's Claude AI attempted to contact the FBI when faced with certain scenarios. This reveals both the sophisticated reasoning capabilities and potential unpredictable behaviors of advanced AI systems.

OpenAI filed a formal legal response to mental health-related lawsuits, outlining fifteen major legal defense strategies. The defense approach is expected to be adopted by other AI companies facing similar litigation. The case highlights growing concerns about AI's impact on mental health and user safety.

Snowflake and Anthropic announced a multi-year, $200 million partnership to bring Claude AI to over 12,600 Snowflake customers. This integration aims to help businesses get accurate answers from their enterprise data while maintaining security.

In December 2024, xAI raised an additional $6 billion in a Series C funding round to achieve an approximately $45 billion valuation. The funds will help accelerate research and development, enhance infrastructure, and bring xAI's products to market. This follows their previous $6 billion raise in May, showing incredible momentum for Elon Musk's AI venture.

Venture capitalist Mitchell Green criticized the massive spending rates at AI companies like OpenAI and Anthropic, calling their burn rates 'ludicrous.' The warning comes amid growing concerns about an AI bubble and questions about the sustainability of current AI investment levels. The comments highlight investor anxiety about AI companies' path to profitability.

Nvidia completed its acquisition of Run:ai, an Israeli startup that helps manage and optimize AI hardware infrastructure. As part of the merger, Run:ai said its software will be open sourced, meaning Nvidia rivals like AMD and Intel will be able to adapt it for their hardware. The deal overcame regulatory hurdles from both EU and US authorities.

OpenAI failed to keep its ChatGPT training logs secret in a copyright lawsuit, potentially exposing how the company trained its models on copyrighted content. This could set important precedents for AI transparency and copyright law.

MIT researchers created VideoCAD, an AI tool that can learn to use computer-aided design software to generate 3D objects from simple sketches. The system could significantly boost designer productivity and help train engineers learning CAD software. The breakthrough represents a major advance in AI-assisted design workflows.

AI investments surged 62% to $110 billion in 2024 while startup funding overall declined 12%. Databricks raised $10 billion, compared to OpenAI's $6.6 billion. Yet, with more than $20 billion to date and another $40 billion reportedly in the works, OpenAI has come to represent a bellwether in the industry.

Google unveiled Gemini 3 and Gemini 3 Pro with claimed 72% answer accuracy and record-breaking reasoning benchmarks. The models are being integrated into Google Search and Workspace, intensifying competition with OpenAI.