• AI KATANA
  • Posts
  • AI-deepfaked meeting tricks Hong Kong employee into handing over $25 million

AI-deepfaked meeting tricks Hong Kong employee into handing over $25 million

Also: Inside OpenAI’s Plan to Make AI More ‘Democratic’

Welcome!

In today’s edition, we delve into a diverse range of AI advancements and applications, from the alarming use of AI-deepfaked video conferences that tricked a Hong Kong employee into transferring $25 million, to OpenAI's initiative to democratize AI development through public deliberation. We explore the transformative impact of AI at the edge in industries like manufacturing and healthcare, highlight an AI-driven breakthrough in decoding Ancient Roman texts buried by Mount Vesuvius, and consider the role of AI in creating virtual dream homes amidst a real estate crisis. Additionally, we cover Google's impressive foray into AI image generation with ImageFX, the integration of ChatGPT into Apple's Vision Pro for augmented reality experiences, and the venture into foundational AI models for the natural world by Jua, a Swiss startup that recently raised $16 million.

Sliced:

  • ⚠️ AI-deepfaked meeting tricks Hong Kong employee into handing over $25 million

  • 🫡 Inside OpenAI’s Plan to Make AI More ‘Democratic’

  • 🦾 Embracing the future with AI at the edge

  • 📜 An Ancient Roman Scroll on Pleasure Was Just Decoded Using AI

  • 🏠 How A.I. Is Remodeling the Fantasy Home

In Hong Kong, a multinational company employee was tricked into transferring $25 million during a video conference call faked using AI, impersonating colleagues. The fraud involved the employee executing 15 transactions to local bank accounts, under the belief they were following legitimate instructions from what seemed to be the company’s CFO and other colleagues. The scam, highlighted by authorities, showcases the sophisticated use of deepfake technology to mimic voices and appearances, underscoring the emerging threats of AI-generated fraud in multi-participant online meetings and the need for heightened vigilance.

OpenAI is pioneering a project to democratize AI by incorporating public deliberation into its AI systems' development and governance. Collaborating with innovators like Colin Megill, who developed Polis for public consensus finding, OpenAI launched a $1 million initiative to explore democratic processes for guiding AI's behavior. Despite challenges, including the risk of AI biases and the complexity of representing diverse public opinions, this initiative represents a step towards engaging the public in shaping AI technologies. The project also reflects on broader questions about the balance of power in AI governance and the potential for AI to support more inclusive decision-making processes.

AI at the edge is revolutionizing industries by decentralizing processing to enhance operational efficiency, reduce latency, and foster innovation. This approach, emphasizing real-time decision-making closer to data sources, is pivotal for sectors including manufacturing, retail, healthcare, and smart cities. For enterprises, the adoption of edge AI is not just beneficial but essential for competitive advantage, necessitating a focus on scalability, security, interoperability, and data governance to navigate the challenges and leverage the opportunities presented by this technological shift.

Ancient Roman scrolls, buried by the eruption of Mount Vesuvius in AD 79, have been partially decoded using artificial intelligence. This breakthrough was achieved by participants of the Vesuvius Challenge, uncovering a text that delves into Epicurean philosophy, specifically discussing pleasure and its relation to scarcity. The winning team, comprising individuals with backgrounds in computer science, machine learning, and robotics, shared a grand prize for their contributions to uncovering these never-before-seen texts from antiquity. This discovery opens new avenues for understanding ancient philosophies and showcases the potential of AI in historical research.

AI is reshaping the concept of homeownership and interior design by offering virtual, highly detailed dream homes that cater to individual tastes, yet are unattainable in reality due to current economic challenges. Through platforms like Instagram and design apps, AI-generated homes provide a form of escapism from the real estate crisis, allowing people to indulge in the fantasy of perfect, personalized spaces without the financial burden. This trend reflects broader societal shifts in how we engage with the idea of home and success, emphasizing virtual satisfaction over tangible achievements.

🛠️ AI tools updates

Google's ImageFX AI image generator, described as a late but impressive entry into the AI image generation space, offers a free, easy-to-use platform that produces high-quality images capable of handling complex prompts with realism. The tool, accessible via Google Labs, includes features like expressive chips for tweaking image details and incorporates a watermarking system, SynthID, by Google DeepMind to identify AI-generated images subtly. It's recommended for a variety of uses, from casual experimentation to more specific applications, with a note on considering legal implications for use in professional work.

OpenAI has launched a ChatGPT app for Apple's Vision Pro, blending AI with augmented reality (AR) in an innovative way. While currently offering basic interaction through voice, the potential for future applications that fully utilize the Vision Pro's AR capabilities is intriguing. This development opens possibilities for more immersive AI interactions, such as generating recipes from ingredients seen or explaining complex problems visually, hinting at a future where AI and AR are closely integrated to enhance our interaction with technology.

💵 Venture Capital updates

Jua, a Swiss startup, has secured $16 million in seed funding to create a foundational AI model for the natural world, beginning with weather and climate patterns. Focused on aiding industries like energy, agriculture, insurance, transportation, and government, Jua aims to offer more accurate modeling and forecasting in response to climate change's volatility. With significant data resources and an ambitious goal to build a system that learns physics, Jua is positioning itself at the forefront of combining AI with natural world understanding, targeting efficiency and comprehensive insight across various sectors.

🫡 Meme of the day

⭐️ Generative AI image of the day

Before you go, check out AI's Impact on Chip Stocks and Earnings.