DeepMind’s Next Big Thing: Gemini

Google’s DeepMind lab, known for its AI program AlphaGo that made history by defeating a Go champion, is now working on a new AI system called Gemini. Using techniques from AlphaGo, Gemini aims to eclipse OpenAI’s ChatGPT in capabilities. This development signifies the continuous evolution of AI technologies and the competitive landscape in the AI industry. Read more

The Marriage of Image Recognition and Generation

In the realm of computer vision, AI has shown remarkable capabilities in both identifying and generating images. These two functions, which have traditionally been separate, are now being combined, leading to exciting new applications and use cases.

AI and Ethics: A Pentagon Perspective

The U.S. Defense Department has been working for over a decade to ensure the responsible use of AI. They have adopted ethical principles for using AI and issued a responsible AI strategy and implementation pathway. This highlights the growing importance of ethical considerations in AI development and deployment.

Will AI Take Over Your Job?

As AI technology advances, concerns about its potential impacts on jobs and employment are growing. While AI is a powerful tool, it’s essential to consider its societal implications and how it might reshape the workforce.

Big Tech’s Billion-Dollar Bets on AI

The competitive landscape in AI, particularly in generative AI, is heating up. Major tech companies like Dropbox and AWS are investing heavily in AI-powered products and generative AI initiatives. Salesforce Ventures plans to invest $500 million into startups developing generative AI technologies, and Accenture and PwC have announced plans to invest $3 billion and $1 billion, respectively, in AI. Read more

The Transparency Issue in AI

The technology underpinning some of today’s most popular AI applications is becoming dangerously opaque, according to Meredith Whittaker, president of secure messaging app Signal. She argues that the current power hierarchy in AI is a significant issue, and structural change is needed to address these challenges.

DeepMind’s AI Controls Robots

DeepMind has developed an AI model, RoboCat, that can perform a range of tasks across different models of robotic arms. This development signifies a significant step forward in the field of robotics and AI.

AI Learning from YouTube

CMU Robotics Institute assistant professor Deepak Pathak showcased VRB, an AI system designed to train robotic systems by watching a recording of a human. This innovative approach to AI training could lead to more advanced and capable robotic systems.

AI Regulation in the EU

European regulators are at a crossroads over how AI will be regulated in the region. The European Consumer Organisation (BEUC) has called for urgent investigations into the risks of generative AI, highlighting the need for robust regulatory frameworks for AI.

AI-Powered Features in Vimeo

Vimeo has launched a suite of AI-powered tools designed to help users create scripts, record footage, and remove unwanted disfluencies from recordings. This shows how AI is being integrated into various digital platforms to enhance user experience.

Synthetic Voices and AI

ElevenLabs, an AI-powered platform for creating synthetic voices, has raised $19 million in a new funding round. This development underscores the growing interest in AI-powered voice synthesis technologies.

Turning Audio into Text with AI

French AI startup Gladia has launched a platform that leverages OpenAI’s Whisper transcription model to turn any audio into text in near real time. This application of AI could revolutionize transcription services and make them more accessible and efficient.

Harness Embraces Generative AI

Harness, a startup creating a toolkit to help developers operate more efficiently, has integrated AI into its platform. The AI can automatically resolve build and deployment failures, find and fix security vulnerabilities, and make suggestions to bring cloud costs under control.

Other Machine Learnings

This week saw the Conference on Computer Vision and Pattern Recognition (CVPR) in Vancouver, Canada, where many interesting talks and papers were presented. One of the highlights was Yejin Choi’s keynote about the possibilities, impossibilities, and paradoxes of AI. She addressed the unexpected limitations of today’s most capable models, such as GPT-4’s difficulty with multiplication, and emphasized the importance of understanding the limitations of AI.

AI in the Metaverse

Intel showcased an interesting model, LDM3D, for generating 3D, 360 imagery like virtual environments. This technology could be used to create on-demand environments in the metaverse.

Voice Cloning with AI

Meta released a voice synthesis tool called Voicebox that’s excellent at extracting features of voices and replicating them, even when the input isn’t clean. This technology could have significant implications for voice cloning in the future.

Keywords: AI Weekly Blog, DeepMind, Gemini, AI Ethics, Pentagon, AI Jobs, Big Tech AI Investments, AI Transparency, RoboCat, AI Robotics, AI Regulation EU, Vimeo AI Features, Synthetic Voices, Audio to Text AI, Harness AI, Machine Learning, Metaverse, Voice Cloning