Introduction to AI: Monthly Roundup of Developments

Hello, fellow tech enthusiasts! Welcome to another edition of AI Monthly, where I share with you the latest and greatest news from the world of artificial intelligence. September was a busy month for AI, with many exciting developments and breakthroughs happening across different domains and applications. Let’s dive right in and see what’s going on in AI space!

Image/Video or Computer Vision

One of the most impressive achievements in computer vision this month was the new inpainting feature Vary Region by Midjourney, a startup that specializes in image editing and manipulation. Vary Region allows users to easily erase unwanted objects or regions from an image and replace them with realistic and coherent content generated by AI. The feature uses a deep neural network that learns from millions of images how to fill in the missing parts of an image in a way that matches the style, color, texture and context of the original image. Users can also adjust the level of detail and diversity of the inpainted content, creating stunning results in seconds. Users can adjust the level of detail and diversity of the inpainted content, creating stunning results in seconds.

Another notable development in computer vision was the launch of YouTube’s new AI-backed hum-search feature, which allows users to find videos by humming or singing a tune. The feature uses machine learning to match the user’s input to a large database of songs and videos, and returns the most relevant results. The feature works even if the user is off-key or doesn’t know the lyrics, as long as they can hum or sing the melody. The feature is available on both mobile and desktop versions of YouTube, and supports multiple languages. You can learn more about this development in TechCrunch’s coverage.

Generative Models

Generative models are AI systems that can create new content or data from scratch, such as images, text, music, etc. One of the most popular generative models is ChatGPT, a large-scale language model developed by OpenAI that can generate coherent and diverse text on almost any topic. ChatGPT has been used for various applications, such as chatbots, content creation, summarization, translation, etc.

This month, OpenAI announced the launch of a new Canva plugin that integrates ChatGPT into the online graphic design platform. The plugin allows users to generate text for their designs using ChatGPT, such as headlines, captions, slogans, etc. Users can also provide keywords or prompts to guide ChatGPT’s output, or choose from different styles and tones. The plugin aims to make design easier and more fun for everyone, especially for those who struggle with writing or creativity. You can learn more about using the Canva ChatGPT plugin in their help documentation.

Language Models

Language models are AI systems that can understand and generate natural language, such as speech or text. Language models have many applications, such as voice assistants, search engines, translation, etc.

One of the most remarkable achievements in language models this month was the development of a brain-computer interface (BCI) that enabled a woman with severe paralysis from a brainstem stroke to speak through a digital voice synthesizer. The BCI was developed by researchers from Stanford University and Neuralink, a company founded by Elon Musk that aims to create neural implants that can connect humans and machines.

The BCI works by recording brain signals from electrodes implanted in the speech-related areas of the brain, and decoding them into words using an AI algorithm. The words are then converted into speech using a synthetic voice. The BCI allows the woman to communicate at a rate of about 15 words per minute, which is comparable to natural speech.

The BCI is a breakthrough for restoring speech to people who have lost it due to injury or disease, and also opens up new possibilities for human-machine interaction. This represents significant advances in neural interface technology for medical applications.

Digital Assistance

Digital assistants are AI systems that can help users with various tasks or requests, such as booking appointments, ordering food, playing music, etc. Digital assistants are usually accessed through voice or text commands on devices such as smartphones, smart speakers, laptops, etc.

One of the most interesting developments in digital assistance this month was the release of AVIA AI Labs, an initiative by AVIA, a market intelligence and advisory services firm for digital transformation in healthcare. AVIA AI Labs aims to harness the power of AI to provide practical examples of its transformative capabilities to the healthcare industry.

AVIA AI Labs consists of three modules: Prompt, Fine-Tune and Monitor. Prompt allows users to generate text for various healthcare scenarios using ChatGPT; Fine-Tune allows users to customize ChatGPT’s output according to their preferences and needs; and Monitor allows users to evaluate ChatGPT’s performance and reliability using metrics and feedback. AVIA AI Labs is designed to help healthcare professionals and organizations explore the potential of AI for improving healthcare quality, efficiency and outcomes.

AI-capable IoT Devices

IoT devices are devices that can connect to the internet and communicate with other devices or systems, such as smart watches, smart thermostats, smart cameras, etc. AI-capable IoT devices are IoT devices that can perform AI tasks, such as recognition, prediction, optimization, etc.

One of the most innovative AI-capable IoT devices that was launched this month was a ‘brainless’ soft robot that can navigate complex obstacles without human or computer direction. The robot was created by researchers from Cornell University and the University of Pennsylvania, and consists of a soft silicone body with four legs that can move independently.

The robot does not have a brain or a central controller, but instead relies on a simple circuit that uses feedback from sensors to generate rhythmic patterns of movement. The robot can adapt to different environments and situations by changing its gait, speed and direction. The robot can also recover from damage or deformation by self-healing its body.

The robot is a demonstration of how AI can be embedded into physical materials and structures, creating new possibilities for robotics and engineering. You can watch the robot in action here: https://www.youtube.com/watch?v=Z8Xx7LjAMgk

The Future of AI

AI is advancing rapidly, and we can expect to see more amazing developments and applications in the near future. Some of the trends and challenges that will shape the future of AI are:

• Ethics and trust: As AI becomes more powerful and pervasive, it also raises ethical and social issues, such as privacy, fairness, accountability, transparency, etc. How can we ensure that AI is aligned with human values and interests, and that it does not cause harm or injustice? How can we build trust and confidence in AI systems and their decisions? How can we regulate and govern AI in a responsible and inclusive way?
• Collaboration and integration: As AI becomes more capable and diverse, it also creates opportunities for collaboration and integration across different domains and disciplines. How can we leverage the complementary strengths of humans and machines, and create synergies between them? How can we combine different types of AI systems and models, such as vision, language, reasoning, etc., to create more intelligent and robust solutions? How can we integrate AI with other technologies, such as biotechnology, nanotechnology, quantum computing, etc., to create new possibilities and innovations?
• Creativity and innovation: As AI becomes more generative and expressive, it also challenges our notions of creativity and innovation. How can we foster human creativity in the age of AI, and avoid being replaced or outperformed by machines? How can we harness AI’s creativity for good, and avoid misuse or abuse? How can we evaluate and appreciate AI’s creativity, and learn from it?

These are some of the questions that will shape the future of AI, and I hope you will join me in exploring them in the next editions of AI Monthly. Until then, stay curious and keep learning!

Sources:

• YouTube Hum Search Feature Coverage - TechCrunch
• Canva ChatGPT Integration Guide
• Soft Robot Demo Video