Weekly AI News of June Third Week 2024

Empowering Defenders through OpenAI’s Cybersecurity Grant Program

OpenAI’s Cybersecurity Grant Program is making waves in the field of cybersecurity. Launched in 2023, this program aims to equip cyber defenders with cutting-edge AI models and foster groundbreaking research at the intersection of cybersecurity and artificial intelligence. Selected projects include:

1. Wagner Lab (UC Berkeley): Pioneering techniques to defend against prompt-injection attacks in large language models (LLMs).
   
   
2. Coguard: Using AI to reduce software misconfiguration, a common cause of security incidents.
   
   
3. Mithril Security: Fortifying inference infrastructure for LLMs, ensuring data privacy even from administrators.
   
   
4. Gabriel Bernadett-Shapiro: Creating AI OSINT workshops and tools for investigators and students.
   
   
5. Breuer Lab (Dartmouth): Developing defense techniques against attacks on neural networks without compromising accuracy.
   

Official Source:-

https://openai.com/index/empowering-defenders-through-our-cybersecurity-grant-program/

Top Things to Know about Copilot+ PCs from Microsoft Surface

Microsoft’s new Copilot+ PCs (Surface Laptop and Surface Pro) are designed to enhance your productivity and creativity. Here’s what you need to know:

1. Powerful Performance: Copilot+ PCs are the fastest, most intelligent Windows PCs on the market. With Snapdragon X Series Processors, they offer faster performance and all-day battery life, thanks to a powerful Neural Processing Unit (NPU) for AI experiences.
   
   
2. Sleek Design: Thin, lightweight, and ultraportable, these devices come in four stunning colors: classic Black, timeless Platinum, bold Sapphire, and refreshing Dune.
   
   
3. Immersive Displays: The new Surface Pro features an OLED with HDR display for a cinematic experience, while the Surface Laptop boasts an HDR touchscreen display with razor-thin bezels.
   
   
4. Everyday AI Companion: The Copilot key gives you easy access to the Copilot app, making AI assistance just a click away.
   

Official Source:-

https://blogs.windows.com/devices/2024/06/18/top-things-to-know-about-copilot-pcs-from-microsoft-surface-available-today-at-microsoft-com/?ocid=FY24_soc_omc_br_x_CopilotPC

Claude 3.5 Sonnet | Advancing AI-Language Models

Anthropic’s Claude 3.5 Sonnet is making waves in the AI world. Here’s what you should know:

1. Frontier Intelligence: Claude 3.5 Sonnet outperforms its predecessor, Claude 3 Opus, in graduate-level reasoning, knowledge, and coding proficiency. It operates at twice the speed, making it ideal for complex tasks like context-sensitive customer support and code orchestration.
   
   
2. State-of-the-Art Vision: It excels in visual reasoning, surpassing Claude 3 Opus on standard vision benchmarks. From interpreting charts to transcribing text from imperfect images, Claude 3.5 Sonnet shines.
   
   
3. Artifacts Feature: Claude 3.5 Sonnet introduces Artifacts—a dynamic workspace where users can interact with AI-generated content in real time. Perfect for integrating AI into projects and workflows.
   

Official Source:-

https://www.anthropic.com/news/claude-3-5-sonnet

DeepMind’s V2A | Soundtracks and Dialogue for Videos

Google DeepMind is developing a new artificial intelligence model that can generate video soundtracks and conversations. The tech giant’s AI research team recently revealed V2A (Video-to-Audio), a new AI model that is still in the early stages of development and “combines video pixels with natural language text prompts to generate rich soundscapes for the on-screen action.”

In the creative community, creating videos based on text prompts is becoming popular. Nevertheless, most of the existing solutions have a significant limitation and can only be used with silent videos.

Google’s text-to-video paradigm, Veo, which was unveiled at I/O 2024 last month, is intended to function flawlessly with Google DeepMind’s V2A.

With this combo, users can enhance the audio quality of their videos in addition to their visual appeal. Additionally, V2A can revitalize “traditional footage” like silent films and archival material, reported the Indian Express.

The goal of this AI model technology is to transform the way viewers produce and interact with videos created by AI. It can be utilized to include dialogue that fits the video’s tone, dramatic music, and realistic sound effects.

Official Source:-

https://deepmind.google/discover/blog/generating-audio-for-video/