Google’s DeepMind has introduced GenCast, a cutting-edge AI weather prediction model that outperforms the European Centre for Medium-Range Weather Forecasts’ (ECMWF) ENS, widely regarded as the global leader in operational forecasting. A study in Nature highlighted GenCast’s superior accuracy, predicting weather more effectively 97.2% of the time during a comparative analysis of 2019 data.
Unlike earlier deterministic models, GenCast creates a complex probability distribution of potential weather scenarios by generating 50 or more forecasts per instance. This ensemble approach provides a nuanced understanding of weather trajectories, elevating predictive reliability.
Google is integrating GenCast into its platforms like Search and Maps, while also planning to make real-time and historical AI powered forecasts accessible for public and research use. With this advancement, the tech giant aims to revolutionise weather forecasting and its applications worldwide.
Congressman Jay Obernolte has called on federal energy regulators to prioritise co-located data centres directly connected to power plants. The move, he argues, is essential for advancing US AI capabilities while addressing energy demands and grid resilience.
The proposal comes amid debates over co-location’s potential impact on grid reliability. A recent Federal Energy Regulatory Commission (FERC) decision rejected an Amazon-Talen Energy project in Pennsylvania, citing grid concerns. Obernolte urged FERC to expedite decisions on such projects to support US AI innovation.
Energy firms Talen and Constellation have challenged FERC’s recent rulings, highlighting the growing importance of energy policies in AI expansion.
Elon Musk’s social media platform X has introduced Aurora, an advanced image generation tool integrated into its Grok AI assistant. Aurora allows users to create photorealistic visuals and explore imaginative concepts. However, some users noted the tool briefly disappeared after its launch.
Aurora, accessible through X’s mobile and web apps, appears to have minimal content restrictions. It can generate images of public and copyrighted figures, though explicit and graphic content is reportedly limited. The tool is still in beta, with Musk promising rapid improvements. While Aurora excels in landscapes and still-life depictions, it struggles with more complex details, like human hands, a common challenge for AI-generated visuals.
The release follows X’s decision to make Grok free for all users, enabling broader access to AI-driven features. Meanwhile, Musk’s xAI team, which developed Aurora, recently secured $6B in funding and is working on further innovations, including Grok 3 and a standalone app.
UCLA is breaking new ground with an AI-developed comparative literature course set to launch in winter 2025. The class, covering literature from the Middle Ages to the 17th century, will feature a textbook, assignments, and teaching assistant (TA) resources generated by Kudu, an AI-powered platform founded by UCLA physics professor Alexander Kusenko. This initiative marks the first use of AI-generated materials in UCLA’s humanities division.
Professor Zrinka Stahuljak, who designed the course, collaborated with Kudu by providing lecture notes, PowerPoint slides, and videos from previous classes. The AI system produced the materials within three to four months, requiring just 20 hours of professor involvement. Kudu’s platform allows students to interact with course content through questions answered strictly within the provided material, ensuring focused and accurate responses.
By streamlining material creation, the approach frees up professors and TAs to engage more closely with students while maintaining consistency in course delivery. UCLA hopes this innovative method will enhance the learning experience and redefine education in the humanities.
The International Committee of the Red Cross (ICRC) has introduced principles for using AI in its operations, aiming to harness the technology’s benefits while protecting vulnerable populations. The guidelines, unveiled in late November, reflect the organisation’s cautious approach amid growing interest in generative AI, such as ChatGPT, across various sectors.
ICRC delegate Philippe Stoll emphasised the importance of ensuring AI tools are robust and reliable to avoid unintended harm in high-stakes humanitarian contexts. The ICRC defines AI broadly as systems that perform tasks requiring human-like cognition and reasoning, extending beyond popular large language models.
Guided by its core principles of humanity, neutrality, and independence, the ICRC prioritises data protection and insists that AI tools address real needs rather than seeking problems to solve. That approach stems from the risks posed by deploying technologies in regions poorly represented in AI training data, as highlighted by a 2022 cyberattack that exposed sensitive beneficiary information.
Collaboration with academia is central to the ICRC’s strategy. Partnerships like the Meditron project with Switzerland’s EPFL focus on AI for clinical decision-making and logistics. These initiatives aim to improve supply chain management and enhance field operations while aligning with the organisation’s principles.
Despite interest in AI’s potential, Stoll cautioned against using off-the-shelf tools unsuited to specific local challenges, underscoring the need for adaptability and responsible innovation in humanitarian work.
Jack Ma, co-founder of Alibaba, made a rare public appearance on Sunday, expressing optimism about the future of Ant Group, the fintech affiliate he also helped establish. Speaking at Ant’s 20th-anniversary celebration, Ma highlighted the transformative potential of AI, stating that the changes driven by AI in the next two decades will surpass current expectations. His remarks, reported by Chinese media outlet 36kr, marked a notable return to the spotlight following his retreat from public life amid regulatory challenges.
Reflecting on Ant Group’s turbulent journey, Ma acknowledged the value of criticism and encouragement in fostering the company’s growth. Ant, the operator of China’s leading mobile payment app Alipay, faced a regulatory crackdown after Ma’s public critique of Chinese regulators in 2020. This led to the cancellation of Ant’s $300 billion IPO, followed by a stringent overhaul of its operations to align with financial regulations. The reforms included Ma relinquishing control of the company in 2023.
Despite these challenges, Ant is charting a path forward, underscored by a leadership transition announced Sunday. President Cyril Han will succeed Eric Jing as CEO starting March 1, 2024. Ma’s renewed confidence in Ant’s potential, especially in the AI era, signals a fresh chapter for the fintech giant as it emerges from years of regulatory scrutiny.
X, owned by Elon Musk, is now offering its AI chatbot, Grok, for free. Users can send up to 10 prompts every two hours and generate ten images during the same period without subscribing. However, certain features, such as analysing more than three images per day, still require a paid subscription.
Previously available only to X Premium members for $8 monthly or $84 annually, Grok’s transition to a freemium model brings it in line with AI offerings like OpenAI’s ChatGPT. The shift follows recent trials of the free version in countries such as New Zealand.
The freemium move coincides with a significant milestone for Grok’s parent company, xAI, which recently raised $6B, bringing its total funding to $12B. With its updated accessibility, Grok aims to broaden its appeal while remaining competitive in the evolving AI market.
The US government has authorised the export of advanced AI chips to a Microsoft-operated facility in the United Arab Emirates. This approval comes as part of Microsoft’s $1.5 billion partnership with Emirati AI firm G42, where the US tech giant holds a minority stake and a board seat. G42 uses Microsoft’s cloud services to support its AI applications.
Concerns arose over potential risks of US AI technology being transferred to China, prompting scrutiny from lawmakers. They sought clarity on G42’s connections to Chinese authorities before permitting the deal to proceed. The export licence requires strict compliance measures, ensuring restricted access to the UAE facility by individuals or organisations from nations under US arms embargoes, including China.
AI-related national security risks, such as the facilitation of weapons development, remain a key issue for US officials. The Biden administration has implemented regulations requiring major AI developers to share system details with the government. G42 has publicly stated its commitment to aligning with international standards in collaboration with US partners and the UAE government.
Ownership ties also add complexity, with G42 partly owned by Abu Dhabi’s sovereign wealth fund and chaired by Sheikh Tahnoon bin Zayed Al Nahyan, the UAE’s national security advisor. The deal underscores a delicate balancing act as Washington navigates strategic and economic interests in the AI sector.
AI startup Perplexity has expanded its publisher partnerships, adding media outlets such as the Los Angeles Times and The Independent. These new partners will benefit from a program that shares ad revenue when their content is referenced on the platform. The initiative also provides publishers with access to Perplexity’s API and analytics tools, enabling them to track content performance and trends.
The program, launched in July, has attracted notable partners from Japan, Spain, and Latin America, including Prisa Media and Newspicks. Existing collaborators include TIME, Der Spiegel, and Fortune. Perplexity highlighted the importance of diverse media representation, stating that the partnerships enhance the accuracy and depth of its AI-powered responses.
Backed by Amazon founder Jeff Bezos and Nvidia, Perplexity aims to challenge Google’s dominance in the search engine market. The company has also begun testing advertising on its platform, seeking to monetise its AI search capabilities.
Perplexity’s growth has not been without challenges. It faces lawsuits from News Corp-owned publishers, including Dow Jones and New York Post, over alleged copyright violations. The New York Times has also issued a cease-and-desist notice, demanding the removal of its content from Perplexity’s generative AI tools.
Cohere, a Canadian AI startup valued at $5.5 billion, is shifting its focus to developing customised AI models for businesses. Co-founder Nick Frosst explained that enterprise users prefer models tailored to specific use cases rather than larger, general-purpose ones. The company aims to refine its approach by prioritising model deployment and customisation over simply increasing model sizes.
Although Cohere will continue building foundation models, it plans to invest in training techniques to improve functionality. The startup has secured over $900 million in funding from major investors like Nvidia, Cisco, and Innovia Capital. Unlike some competitors, Cohere positions itself as an independent player, working with clients such as Oracle and Fujitsu to design models for their unique requirements.
The AI industry, once focused on scaling up models, now faces diminishing returns from increasing model size. As large language model advancements plateau, Cohere’s customised approach offers a more efficient and cost-effective solution. Frosst highlighted that this strategy aligns with the company’s enterprise-centric vision and avoids reliance on speculative breakthroughs in artificial general intelligence.
By concentrating on tailored AI solutions, Cohere aims to enhance real-world applications for its enterprise clients. This strategy positions the startup as a competitive alternative to larger AI labs such as OpenAI and Anthropic.
