Google Cloud's AI Revolution: Insights from CEO Thomas Kurian

Born Digital Summit Unveils Google Cloud's AI Endeavors

In a significant revelation at the Born Digital Summit held in Bengaluru, Thomas Kurian, CEO of Google Cloud, provided insights into the company's extensive AI efforts and its ongoing mission to bolster global infrastructure.

With a history of AI engagement dating back to 2004, Google Cloud now boasts two data regions in India, located in Delhi and Mumbai, respectively.

Google Cloud's AI Solutions: Vertex and Duet

Google Cloud is currently at the forefront of AI innovation, offering two primary AI solutions: Vertex for developers and Duet for end-users and businesses. An astounding achievement is Google Cloud's management of approximately 34 percent of global internet traffic.

Kurian proudly highlighted that Google Cloud can train an AI model twice as fast and at half the cost compared to its competitors.

A Diverse Model Garden

Kurian revealed that the Google Cloud Model Garden comprises a diverse array of over 100 models, catering to various needs. These models encompass text, chat, programming code, images, audio, and visuals.

What sets Google Cloud apart is its commitment to continuous improvement, with model upgrades every two weeks and generational upgrades every four months.

Real-World Adoption and Impact

Google Cloud's AI models have not remained confined to research and development. Currently, a staggering 24,000 developers within Google utilize the Google Cloud AI-powered coding model. Impressively, over 75 percent of them accept nearly all instructions. Notable collaborations include Moderna, a key player in the development of COVID-19 vaccines, utilizing Duet AI to expedite drug simulation through mathematical models.

AI Empowering Everyday Applications

Thomas Kurian, CEO of Google Cloud emphasized the integration of AI into everyday applications. Gmail on smartphones, for instance, leverages a compact AI model that operates locally, ensuring super-low latency. However, larger models, while versatile, are inherently slower due to the evaluation of numerous parameters.

Cortex AI: Enhancing User Interactions

Through the Cortex AI platform, Google Cloud offers natural language-based interactions for retrieving information and conducting transactions, enhancing user experiences and accessibility.

Battling Hallucination Effects with "Grounding"

Acknowledging concerns related to generative AI, Google Cloud is actively working on reducing hallucination effects. The introduction of "grounding" ensures fresh results, accompanied by fact-checks, limiting inaccuracies.

Vision Watermarking: Copyright Protection

In a groundbreaking feature, Google Cloud AI introduces vision watermarking, enabling the identification of AI-generated images even when they lack watermarks. This innovation addresses copyright concerns surrounding AI-generated content.

Unveiling PaLM and Med-PaLM

Google recently unveiled the latest generation of the PaLM (Pathways Language Model) model, capable of summarizing entire books. Additionally, Med-PaLM is a specialized AI model designed to provide medicine-related results.

Empowering End-Users

Google Cloud's Duet-AI-powered "Help Me Write" feature on Gmail has gained significant traction, with over a million users since its launch in March 2023. Furthermore,

Google Sheets now benefits from an AI-powered analyst capable of generating formulas and executing calculations. Google Slides introduces the capability to generate photo-realistic images with text inputs.

As Google Cloud continues to push the boundaries of AI innovation, these developments hold the promise of transforming industries and enhancing user experiences across the digital landscape.