Fei Fei Li “Godmother of AI” Outlines 3 Fundamental Principles for AI Policy Ahead of AI Action Summit in Paris

⚡️ Headlines

🤖 AI

China's DeepSeek AI Chatbot Challenges U.S. in Complex Reasoning Tasks – The rise of China's DeepSeek AI signals increasing competition in advanced AI reasoning capabilities. [Washington Post]

How DeepSeek AI's News Citation Approach Stacks Up Against Rivals – DeepSeek faces scrutiny for how it sources and cites news content compared to other AI systems. [Nieman Lab]

China’s DeepSeek AI Rise Fuels Talent Battle with U.S. – As China’s AI sector grows, DeepSeek drives demand for domestic AI talent and challenges U.S. dominance. [Rest of World]

China's DeepSeek AI Raises Alarm for Potential Misinformation Risks – Experts warn of DeepSeek's potential to generate dangerous or false information. [Wall Street Journal]

AI Critics Urge Christie's to Cancel Upcoming AI Art Auction – An open letter calls on Christie’s to stop an auction featuring AI-generated artwork due to ethical concerns. [Open Letter]

Exclusive: OpenAI Co-Founder Sutskever Sees Breakthroughs in General AI by 2026 – Ilya Sutskever predicts imminent advances in general AI, bringing it closer to human-level understanding. [Yahoo Finance]

🦾 Emerging Tech

Trump-Endorsed Crypto Memecoin Sparks Market Frenzy – A Trump-endorsed memecoin gains unexpected traction in crypto markets, raising regulatory concerns. [New York Times]

Circle's Stablecoin Rebounds After 2024 Crypto Market Collapse – Circle’s stablecoin regains lost market share following last year’s industry-wide crypto crash. [Bloomberg]

🔬 Research

Optimizing Transformer Models for Enhanced Long-Range Attention – New research proposes more efficient transformer architectures to improve long-range sequence modeling. [arXiv]

Advances in Quantum Error Correction Using Surface Codes – This study introduces methods for improving quantum error correction in surface code systems. [arXiv]

⚖ Legal

Leaked AI Action Summit Document Outlines Urgent Regulatory Proposals – A leaked statement reveals plans for stricter AI regulations discussed at a closed-door AI summit. [Transformer News]

🎱 Random

U.S.-China Tech Trade War Intensifies as New Restrictions Loom – The ongoing U.S.-China tech trade conflict escalates with fresh restrictions on critical technology exports. [Wall Street Journal]

🔌 Plug-Into-This

AI Pioneer Fei-Fei Li Advocates for Science-Based AI Policy

In a recent op-ed, Fei-Fei Li, co-director of Stanford's Institute for Human-Centered AI, emphasizes the need for AI policies grounded in empirical science rather than speculative scenarios. She outlines three guiding principles: basing policy on current scientific understanding, adopting pragmatic approaches to minimize unintended consequences while fostering innovation, and empowering the AI ecosystem through open access and collaboration.

• Science Over Speculation: Li stresses that effective AI governance should rely on empirical data and rigorous research, avoiding distractions from futuristic, speculative scenarios.

• Pragmatic Policy Development: She advocates for policies that are practical and minimize unintended consequences, such as exacerbating biases, while promoting innovation.

• Empowering the AI Ecosystem: Li highlights the importance of open access to AI models and tools, emphasizing that collaboration among academia, open-source communities, and the private sector is crucial for progress.

• Role of Established Institutions: She suggests that organizations like the U.S. National Institute of Standards and Technology could provide accurate, up-to-date information about AI's real-world effects, leading to precise, actionable policies grounded in technical reality.

• Balancing Innovation and Risk Mitigation: Li acknowledges the dual potential of AI to democratize access to high-quality medical care and to exacerbate existing biases, advocating for governance that tactically mitigates risks while rewarding responsible implementation.

🧩 Li's call for science-based AI policy aligns with a broader movement among experts advocating for responsible AI development, emphasizing the need for policies that are both innovative and ethically sound.

Christie's Announces AI Art Auction Amid Controversy

Christie's has unveiled plans for the "Augmented Intelligence" auction, featuring 20 AI-generated artworks with prices ranging from $10,000 to $250,000. This announcement has sparked significant backlash from the artistic community, with thousands of artists urging the auction house to cancel the event due to concerns over the use of AI in art creation.

• Artists' Concerns: Over 3,000 artists have signed an open letter claiming that the AI models used to create these artworks are trained on copyrighted works without permission, constituting "mass theft." They argue that this practice exploits human artists by using their work without authorization or compensation.

• Christie's Response: The auction house asserts that, in most cases, the AI utilized in these artworks has been trained on the artists' own inputs. They emphasize that the featured artists have strong, multidisciplinary practices and are using AI to enhance their existing bodies of work.

• Defending AI-Assisted Art: Artists like Mat Dryhurst and Refik Anadol, whose works are included in the auction, defend their use of AI. Dryhurst views his piece as an exploration of how his wife's concept appears in publicly available AI models, asserting that it is within their rights to engage with this technology. Anadol dismisses the backlash as a result of "lazy critic practices and doomsday hysteria."

• Not the First Time: In 2018, Christie's sold "Edmond de Belamy," an AI-generated portrait, for over $400,000, which also faced criticism for its creation process and the use of uncredited code from another AI artist.

🎨 AI-generated art has been a source of hot debate for years now with the artistic communities being some of the loudest voices of discontent — there’s an ongoing open letter with more than 3000 signatures calling for this auction’s cancellation.

Determining Compensation for Authors Licensing Books for AI Training

As artificial intelligence systems increasingly rely on vast amounts of textual data for training, the question of how much authors should be compensated for the use of their works has become a more pressing issue. While there is a consensus that creators deserve payment, establishing appropriate compensation models remains complex.

• Consensus on Compensation: Experts agree that tech companies should remunerate creators whose works are utilized in AI training.

• Challenges in Valuation: Determining the value of individual works in the context of AI training is challenging due to the vast and varied nature of the data sets involved.

• Proposed Frameworks: Some scholars suggest frameworks that compensate copyright owners proportionally to their contributions to AI-generated content, using techniques from cooperative game theory to assess the impact of specific works on the AI's outputs.

• Legal Precedents: The Writers Guild of America secured an agreement ensuring that AI will not replace human writers, highlighting the industry's recognition of the importance of protecting creators' rights in the face of advancing technology.

• Ongoing Negotiations: Industry players and creative organizations are working to establish standardized licensing agreements that balance fair compensation with the growth of AI innovations.

📚 The debate signals a shift toward formalizing the relationship between creative industries and AI development, with outcomes that could shape compensation norms across media sectors.

 🆕 Updates

📽️ Daily Demo

🗣️ Discourse