The SEC Newgate AI Weekly

In this week’s AI Weekly, you'll gain insight into the latest developments from Google's I/O 2024, where Project Astra spearheads a revolution in AI virtual assistants, and discover the latest advancements with Gemma 2 and PaliGemma models, promising transformative capabilities in image processing and comprehension. We also explore OpenAI's GPT-4o release as an 'emotional' AI chatbot. Amidst the tech excitement, we also confront the environmental consequences of AI, urging sustainable practices and explore the depths of the ocean's mysteries as researchers employ machine learning to decode the language of sperm whales. 

Google's bold moves at I/O 2024 

At Google's annual I/O conference, attendees were treated to an exciting glimpse into the future of the Google ecosystem. Among the highlights was Project Astra, a pioneering initiative redefining AI assistant. It seamlessly processes multimodal information, querying real-time data through your phone's camera and receiving natural conversational responses. You can watch a demo of this in action here. 

Google also unveiled significant updates across its Gemini family of models, introducing the next generation of open models - Gemma 2 and PaliGemma. Google showcased their capabilities in image and short video captioning, visual question answering, text comprehension within images, object detection, and object segmentation. These advancements have the potential to transform how we interact with technology, opening new possibilities for innovation and convenience.  

OpenAI releases “emotional” AI chatbot  

In case you missed it - OpenAI announced on Monday the launch of its latest AI model - GPT–4o. This is its smartest model yet and has the ability to laugh, listen, look and talk. My colleague Matt Redley has written an informative piece on everything you need to know.  

The environmental consequences of AI 

Amidst the flurry of new tools being released by industry giants, it's crucial to take stock of the broader impact of AI and data centres on our planet's resources. Particularly, the strain on our water supply and the surge in carbon emissions into the atmosphere. Microsoft recently reported a significant 31% increase in emissions stemming from its supply chain and other indirect sources in 2023. The expansion of infrastructure to support the burgeoning demands of generative AI operations emerged as a primary driver behind this rise. These findings underscore the pressing need for sustainable practices and innovative solutions as we explore these tools. 

Deciphering sperm whale conversations with machine learning 

On a lighter note, away from the bustling tech giants, researchers at MIT CSAIL and Project CETI have embarked on an extraordinary journey into the depths of the ocean's mysteries to unlock the language of sperm whales with the helping hand of machine learning technologies. While the ocean's chorus has fascinated scientists for decades, this new research promises a fresh perspective, revealing layers of complexity previously unseen in the conversations of these majestic sea dwellers.