Google I/O 2024 Unveils Extensive AI Dominance: Gemini AI Introduces ChatGPT 4o Features
The Google I/O 2024 developer conference was held as scheduled, with a focus on the latest artificial intelligence technology and product updates. In the two-hour keynote speech, Google unveiled a series of innovative technological breakthroughs and updates aimed at providing developers and consumers with more convenience and innovative experiences.
Table of Contents:
Toggle
Next-generation development tool: Firebase Genkit
AI applications are the focus
AI innovation in education: LearnLM
New feature for educational videos on YouTube: AI-generated quizzes
AI enhancements: Updates to Gemma 2
New discovery feature in Google Play
Protecting user safety: Detecting fraudulent activities during calls
Innovative search and interaction: Ask Photos
Gemini AI applications
Gemini in Gmail
Gemini 1.5 Pro: Doubling processing power
Gemini Live: Real-time interactive experience
Gemini Nano: Chrome-integrated micro AI
Gemini applications on Android
Gemini applications in Google Maps
Expanding AI capabilities: Performance improvement of Tensor Processing Units
At this year’s conference, Google introduced a new platform called Firebase Genkit. This open-source framework is designed to simplify the development of artificial intelligence applications using JavaScript/TypeScript and will soon support the Go language. The launch of Firebase Genkit aims to accelerate the implementation of AI features in both new and existing applications, covering various purposes such as content generation, summarization, text translation, and image generation.
In this speech, Google mentioned AI technology 121 times, demonstrating the company’s deep involvement in this field. CEO Sundar Pichai even emphasized that this is part of Google’s effort to showcase its AI initiatives to the outside world.
Google also introduced a new generation of AI models called LearnLM, specifically designed for “training” in the education sector. Developed in collaboration between Google’s DeepMind AI research division and Google Research, this series of models aims to support student learning through interactive teaching. LearnLM has been tested in Google Classroom and will be further integrated into curriculum planning and optimization processes.
YouTube has added an AI-generated quiz feature, allowing viewers to interact while watching educational videos, such as asking questions, getting explanations, or participating in quizzes. This feature provides a new learning experience for users who need to watch long educational videos.
In response to the needs of developers, Google will soon add a new model with 2.7 billion parameters to Gemma 2, optimized by Nvidia for efficient operation on the next generation of GPUs.
Google Play has updated its discovery feature, making it easier for users and developers to promote and find applications. This includes new user acquisition methods, updates to Play Points, and enhancements to tools and APIs for developers.
Google previewed a new feature that can detect potential fraudulent activities in calls in real-time, which will be integrated into future versions of Android. By analyzing the conversation patterns during calls, this system can effectively warn users about security risks.
Google Photos is launching an experimental feature called “Ask Photos,” which uses AI to understand the content and metadata of photos. Users can use natural language queries to search, making the search process more intuitive and requiring less manual intervention.
In Gmail, users will be able to search, summarize, and draft emails using Gemini AI technology. Additionally, Gemini AI can perform more complex tasks such as handling e-commerce returns, including searching the inbox, finding receipts, and filling out online forms.
The upgrade to Gemini 1.5 Pro enables it to analyze longer documents, code repositories, videos, and recordings than before. In the latest private preview, the processing power of this flagship model has increased to handle up to 2 million tokens, twice as much as before.
Google previewed a new feature called Gemini Live, which allows users to have deep voice conversations with Gemini through their smartphones. Users can interrupt Gemini’s questions at any time during the conversation, and the system will adapt in real-time to the user’s voice patterns. Additionally, Gemini can identify and respond to the user’s surrounding environment through the user’s smartphone camera.
Gemini Nano, the smallest member of the Google AI model family, will be directly integrated into the Chrome desktop client starting from Chrome version 126. This will allow developers to implement their own AI features using models on their devices, such as the “Smart Compose” tool in Gmail.
Google’s Gemini AI will replace Google Assistant and be deeply integrated into the Android system, allowing users to directly drag and drop AI-generated images into applications such as Gmail and Google Messages. YouTube users will also be able to use the “Ask this video” feature to search for specific information in videos.
Gemini’s capabilities will be applied to the Google Maps developer platform, starting with the Places API. Developers will be able to showcase location and area summaries analyzed by Gemini in their own applications and websites. Developers will no longer need to write their own custom location descriptions.
Google also announced the sixth generation of its Tensor Processing Units (TPU) AI chips, named Trillium, which will be released later this year with significant performance improvements.
Google I/O 2024 showcased Google’s latest achievements in AI and technological innovation, from educational tools to developer resources, with each update aiming to improve efficiency, enhance interactivity, and ensure user safety.
AI
Gemini AI
Google I/O
