🔐 Google Unveils AI-Powered Security Boost for Workspace with Gemini AI-Powered Advisor

🚀 IBM Unveils Major Expansion of watsonx AI Platform with Meta's Llama 3.2 Models

Author

Ahmad Ak
27 Sep • Estimated Reading Time: 10 minutes

Tech Insights AI 0020, Friday 27th September 2024

🔐 Google Unveils AI-Powered Security Boost for Workspace with Gemini AI-Powered Advisor

Google has announced significant security enhancements powered by Gemini AI for organizations of all sizes, with a focus on improving Gmail protection and introducing a new security advisor tool. These improvements are set to roll out over the next few weeks to paid Google Workspace customers.

Security Advisor for Gmail

  • The new security advisor tool aims to bring enterprise-grade security to organizations of any size, providing tailored intelligence to defend against evolving threats.

    Key features include

  • Security sandbox: All email attachments will be scanned for malicious software in a virtual environment.

  • Enhanced safe browsing: Additional protection by scanning incoming messages for malicious content before delivery.

  • AI-powered filtering: A large language model trained to spot abuse patterns and deploy rapid protections.

Image generated using Flux

The security advisor will deliver actionable guidance directly to IT admins' inboxes, allowing them to customize scan rules and automatically move suspicious messages to the spam folder.

Expanded Protection Beyond Gmail

The security advisor also brings improvements to other Google services:

  • Chrome enhanced safe browsing: Protects users from malware, phishing, and other online threats.

  • Data protection: Prevents unwanted sharing of sensitive information.

  • App access protection: Gives admins management options for access to core Google Workspace apps.

These features will be accessible through the Google Workspace admin console, providing customizable threat defense, account protection, and data safeguards.

Gemini AI Security Enhancements

Google is also implementing enterprise-grade security protections for the Gemini AI app, available from October 15 for customers on Workspace Business, Enterprise, or Frontline plans.

Key aspects include:

  • Subjecting Gemini to the same privacy, security, and compliance policies as other Workspace core services.

  • Allowing Workspace admins to control the storage of Gemini user prompts and generated responses.

  • Ensuring Gemini does not store organization's data, user prompts, or generated responses for AI model training.

Research on Generative AI Usage

Google's recent research on generative AI adoption revealed:

  • 61% of c-suite leaders at major enterprises are already deploying AI tools.

  • 56% of executives reported improved security posture due to generative AI.

  • 82% cited improved threat identification, and 71% reported reduced time to resolve security issues.

🚀 IBM Unveils Major Expansion of watsonx AI Platform with Meta's Llama 3.2 Models

This launch represents a significant milestone in IBM's AI journey, offering businesses powerful new tools to harness the potential of generative AI and vision models in a secure, scalable, and customizable leap forward in IBM's commitment to open-source AI and its mission to provide cutting-edge AI capabilities to enterprises.

Llama 3.2 Models Now Available on watsonx

The integration includes:

o Llama 3.2 90B Vision and 11B Vision: These multimodal models can process both text and images, enabling advanced visual reasoning tasks such as document understanding, chart interpretation, and image captioning.

o Lightweight 1B and 3B Models: Designed for edge and mobile devices, these models offer powerful capabilities while maintaining privacy and reducing latency.

Expanding watsonx Capabilities

With this integration, watsonx.ai now offers

o Enhanced Visual AI: The ability to reason on high-resolution images up to 1120x1120 pixels, supporting tasks like classification, object detection, and OCR.

o Improved Edge Computing: The lightweight models enable on-device AI applications, ensuring data privacy and real-time processing.

o Broader Model Selection: Users can now choose from an expanded library of foundation models, including open-source options like Llama 2 and StarCoder.

IBM's Commitment to Open Source

This launch reinforces IBM's dedication to open-source AI development. The watsonx platform leverages a rich stack of open-source technologies, including:

o CodeFlare for managing distributed AI workloads

o PyTorch for model development and inference

o KServe for serving machine learning models at scale

Impact on Enterprise AI

The integration of Llama 3.2 models into watsonx is expected to accelerate AI adoption in enterprises by:

o Enabling more sophisticated visual AI applications

o Improving privacy and security with on-device processing

o Offering a wider range of models to suit various business needs

Looking Ahead

As AI continues to evolve rapidly, IBM's watsonx platform is positioning itself as a leader in enterprise-ready, open-source AI solutions. The collaboration with Meta and the integration of Llama 3.2 models demonstrate IBM's commitment to staying at the forefront of AI innovation while prioritizing openness, flexibility, and responsible AI development.

This launch represents a significant milestone in IBM's AI journey, offering businesses powerful new tools to harness the potential of generative AI and vision models in a secure, scalable, and customizable environment.

🗣️ Tech Tycoon Slams Microsoft's AI Chatbot

Billionaire executive Ghodsi'sbricks CEO Ali Ghodsi has unleashed a scathing critique of Microsoft's AI chatbot, Claude, calling it a "science project" and questioning its practical utility. In a bold statement that has sent ripples through the tech industry, Ghodsi asserted that Claude is "not useful for anything".

Ghodsi's comments come at a time when AI chatbots are increasingly in the spotlight, with tech giants vying for dominance in this rapidly evolving field. The Databricks CEO's criticism specifically targeted Claude's tendency to hallucinate or generate false information, a common challenge faced by large language models.

Implications for the AI Industry

This high-profile criticism raises important questions about the current state and future direction of AI chatbot technology:

  • Reliability Concerns: Ghodsi's remarks highlight ongoing issues with AI chatbot accuracy and reliability.

  • Practical Applications: The comments challenge the tech industry to demonstrate real-world utility for these AI systems.

  • Competition Heats Up: Ghodsi's critique may intensify the race among tech companies to develop more robust and practical AI chatbots.

📱 Meta and Arm Partner to Bring Advanced AI to Smartphones

Meta and Arm have unveiled a groundbreaking collaboration aimed at integrating powerful AI capabilities into smartphones and other mobile devices. The partnership introduces two personalizedodels, Llama 3.2 1B and 3B, specifically designed to run efficiently on Arm-based mobile chips

Image created using Flux

These new models, with 1 billion and 3 billion parameters respectively, are optimized for on-device processing, enabling faster user experiences and improved energy efficiency. The collaboration aims to revolutionize smartphone functionality, potentially allowing users to interact with their devices through conversational interfaces rather than traditional button presses

With over 300 billion Arm-based devices worldwide, this partnership could significantly accelerate the adoption of AI in mobile technology, paving the way for more intuitive and personalized user experiences across a wide range of devices, from smartphones to smartwatches and even security cameras

🚗 Samsung Unveils New SSD to Power Automotive AI Revolution

Samsung has introduced a groundbreaking solid-state drive (SSD) specifically designed to meet the demanding requirements of artificial intelligence in the automotive industry. The new PM1743A SSD is tailored for in-vehicle infotainment (IVI) systems and advanced driver-assistance systems (ADAS), marking a significant leap forward in automotive technology.

Key features of the PM1743A include:

  • PCIe Gen4 interface for high-speed data transfer

  • Capacities ranging from 256GB to 2TB

  • Enhanced reliability with a mean time between failures (MTBF) of 2 million hours

  • Automotive temperature grade of -40°C to 105°C for extreme conditions

  • AEC-Q100 Grade 3 qualification, ensuring durability in automotive environments

This innovative SSD addresses the growing need for robust, high-performance storage solutions in modern vehicles, capable of handling the massive amounts of data generated by AI-powered systems. Samsung's move positions the company at the forefront of the automotive AI revolution, potentially transforming the driving experience and paving the way for more advanced autonomous vehicles in the future

💻 Tech Snapshots 📸

Zeta Global has launched new AI-driven mobile marketing solutions, including customizable AI Agents, aimed at enhancing personalized campaigns and improving customer interactions. The company's stock (NYSE:ZETA) rose following the announcement, reflecting investor optimism about the potential impact of these advanced tools on marketing efficiency and ROI.

Nokia has been selected by CoreWeave to deploy its IP routing and optical transport portfolios globally, enhancing the hyperscaler's AI and machine learning capabilities. The partnership aims to provide superfast, reliable performance at scale while improving energy efficiency, meeting the growing demands of AI-centered applications and complex workloads

The Rabbit R1, a pocket-sized AI assistant device, has reached 5,000 daily active users according to CEO Jesse Lyu. Despite mixed reviews and limited functionality, the $199 gadget has garnered interest from early adopters, with the company shipping over 100,000 units since its launch earlier this year.

🛠️ Latest AI Tools and Setup Guides

Zapstart.io Validate startup ideas instantly with AI

Machina Eco Sustainable AI chatbot: AI for a world we like to live in.

Nexus AI Your all-in-one AI platform for content creation and design.