• AI KATANA
  • Posts
  • OpenAI, Google DeepMind's current and former employees warn about AI risks

OpenAI, Google DeepMind's current and former employees warn about AI risks

Also: Photo of Jensen Huang goes viral

Author

AI KATANA
June 04, 2024

Morning!

In today's newsletter, we highlight pressing concerns from current and former OpenAI and Google DeepMind employees about AI risks, emphasizing the need for better oversight and regulatory measures. We also cover ASEAN's strategic steps to govern AI through a new guide focused on ethical and responsible use. On the tech front, AMD's new Ryzen AI 300 Series SoCs are set to revolutionize AI PC performance, and Cisco's $1 billion AI fund has made its first significant investments, reflecting a broader trend of tech giants capitalizing on AI advancements. Additionally, Adobe Lightroom's latest AI tools promise to transform photo editing, making advanced techniques accessible to all users.

Sliced:

  • ⚠️ OpenAI, Google DeepMind's current and former employees warn about AI risks

  • 🌏 ASEAN is laying the foundation to effectively govern AI use

  • 🔥 AMD Heats Up The AI PC Race With Ryzen AI 300

  • 💵 Cisco launches $1 billion AI fund and makes first investments

  • 📷 Photo of Jensen Huang goes viral

Current and former employees from OpenAI and Google DeepMind have issued a warning about the risks posed by AI, emphasizing that the financial motives driving these companies hinder effective oversight. Their open letter highlights several potential dangers, including the spread of misinformation, the loss of control over AI systems, and the exacerbation of existing inequalities, which they suggest could ultimately threaten human existence. They criticize the weak obligations of AI firms to share information with governments and call for a better process for employees to voice concerns without being silenced by confidentiality agreements. The letter also notes incidents where AI-generated content has produced voting-related disinformation, underscoring the urgent need for regulatory measures.

ASEAN is taking significant steps to govern the use of AI through the introduction of the ASEAN Guide on AI Governance and Ethics, which aims to help member states develop domestic regulations. This initiative is essential for fostering trust, ensuring responsible and ethical AI usage, and providing a foundation for a shared understanding across the region. The guide emphasizes a nuanced, risk-based approach and the importance of keeping humans involved in AI processes. It encourages iterative risk assessments and outlines clear roles and responsibilities for AI governance. Additionally, ASEAN plans to establish a working group to support regional AI governance policies, highlighting the need for international interoperability and collaboration. This effort aligns with global trends, as seen with the EU's AI Act, and underscores the importance of regulatory cooperation to address emerging AI challenges responsibly.

AMD has introduced its Ryzen AI 300 Series mobile SoCs at Computex 2024, marking a significant advancement in AI PC technology. These new SoCs surpass Microsoft's performance requirements by delivering 50 TOPS of NPU performance, a notable increase over the required 40 TOPS. The Ryzen AI 300 series features up to 12 cores, 24 threads, XDNA 2 NPU, and RDNA 3.5 graphics with 16 compute units. These enhancements promise up to five times the performance of the previous generation NPU and double the power efficiency. Expected to launch in July, the new PCs will support advanced AI applications, including Microsoft's Copilot, and AMD is collaborating with over 150 software developers to optimize AI applications for this platform. The introduction of these SoCs signals a competitive push in the AI PC market, with a focus on improving on-device AI capabilities and performance.

Cisco has launched a $1 billion fund to invest in AI startups, joining other tech giants in seeking stakes in emerging AI companies. The fund, managed by Cisco's investment arm, has already committed nearly $200 million to companies such as Cohere, Mistral AI, and Scale AI. Scale AI, a data startup, is valued at nearly $14 billion, while Cohere and Mistral are reportedly seeking valuations of $5 billion each. This move is part of Cisco's broader strategy, having made over 20 AI-focused acquisitions and investments in recent years, to integrate generative AI and machine learning across its portfolio. This investment comes as tech companies globally are intensifying their efforts to capitalize on the AI boom sparked by the launch of OpenAI's ChatGPT in 2022.

Has Jensen reached rockstar status? 🤩

🛠️ AI tools updates

Cohere, a Toronto-based AI startup, has secured a $450 million investment at a $5 billion valuation from notable investors including Nvidia, Salesforce Ventures, Cisco, and the Canadian pension fund PSP Investments. This new valuation more than doubles Cohere's previous worth from its $270 million Series C round last June. Cohere specializes in developing large language models for AI applications, such as interactive chat and text generation. This investment reflects a broader trend of significant valuation increases for AI startups, as seen with companies like CoreWeave and Mistral AI. Concurrently, Cisco has launched a $1 billion AI investment fund to support the startup ecosystem and has already allocated nearly $200 million to various AI firms, including Cohere, Mistral AI, and Scale AI.

💵 Venture Capital updates

Adobe Lightroom has introduced new AI-powered features to enhance photo editing, making it easier for users to apply advanced edits with just a few clicks. The Generative Remove tool allows users to effortlessly remove unwanted objects or people from photos by using AI to fill in the gaps with matching elements from the surrounding areas. Additionally, the enhanced Lens Blur feature uses AI to create professional-looking background blurs, identifying the main subject and accurately distinguishing between foreground and background. These tools, available in both Lightroom and Lightroom Classic, as well as their mobile apps, are designed to simplify complex edits and improve overall photo quality with minimal effort.

🫡 Meme of the day

⭐️ Generative AI image of the day