The Artificial Intelligence Podcast

Join Arijit Raychowdhury, Chair of Electrical and Computer Engineering at Georgia Tech, and Cheryl Martin, Global Higher Education and Research at Nvidia, as they discuss their pioneering collaboration in AI education. Diving into the genesis of the partnership, they elaborate on how it aims to democratize access to AI tools and resources for students across disciplines. From hands-on experience to mentorship opportunities, they explore the unique setup of the AI makerspace and its potential to inspire the next generation of innovators.

Meta, the parent company of Facebook, is introducing smart assistants powered by artificial intelligence (AI) across its apps, including Instagram, WhatsApp, Messenger, and Facebook. The rollout, powered by Meta's AI model LLaMA 3, aims to bring AI-powered assistance to billions of users in meaningful ways. Users will be able to ask Meta AI for help with tasks and information, and the technology will be integrated into various parts of the apps. Meta's move stands out due to the massive scale of its apps, which are used by nearly four billion people worldwide each month. The AI assistants will be free to use and will enhance user experiences within Meta's apps.

US-based chip design firm AMD has announced its latest processors for AI PCs, aiming to compete against rivals Nvidia and Intel. The Ryzen Pro 8040 series processors for laptops and Ryzen Pro 8000 series processors for desktops are built using advanced 4-nanometer technology, making them highly powerful chips for business PCs. They are set to power PC models from brands such as HP and Lenovo starting in the second quarter of 2024. The demand for AI PCs is predicted to grow, with shipments expected to make up 22% of all PCs by 2024, according to consulting firm Gartner.

Researchers from the University of Cambridge have used artificial intelligence (AI) to accelerate the search for treatments for Parkinson's disease. Using machine learning, they were able to quickly screen millions of compounds in a chemical library and identify five highly potent ones that can block the clumping of the protein alpha-synuclein. This process, which normally takes months or years, was sped up by ten-fold and cost reduced by a thousand-fold using AI. The breakthrough could lead to the development of effective treatments for Parkinson's more quickly, benefiting the millions of people worldwide affected by the disease.

Adobe is working on developing generative AI video tools for its Premiere Pro video editing platform. These tools will allow users to generate and edit videos using text prompts, add or remove objects from videos, and extend the length of video clips. Adobe is also exploring integrations with third-party AI models from companies like Runway, Pika Labs, and OpenAI's Sora, which would provide more options for users. While the release date is not yet announced, Adobe has showcased the capabilities of its own video model. The integration of AI models into Premiere Pro aims to provide creative professionals with a seamless editing experience and more creative possibilities.

The U.S. government is taking steps to regulate AI and ensure society can effectively adapt to its introduction, although the regulatory environment appears relatively lenient compared to the European Union. Klarna, a financial services provider, predicts that its AI assistant tool will increase profits by $40 million by the end of 2024. Open AI's systems, which power Klarna's AI tool, have garnered attention from Congress. The U.S. Senate Task Force on AI has enacted at least 15 bills into law, but concerns remain about the potential impact of AI on job prospects. Some lawmakers have proposed robot taxes to limit job losses, but economists argue for setting them at a relatively low level to avoid hampering technological growth.

The Air Force is exploring the use of artificial intelligence (AI) in wargaming to enhance decision-making and strategic analysis. Traditional wargaming methods can be expensive and limited in analysis, and AI has been identified as a potential solution. Lt. Gen. David Harris of the Air Force Futures office highlighted the use of AI assistants to run thousands of iterations in a single turn of wargaming, allowing for a better understanding of optimal solutions in different conflict scenarios. The Air Force is also interested in adjusting the levels of autonomy for future systems, particularly human-machine teaming, and is pursuing next-generation drones with advanced capabilities.

Researchers from MIT, the Broad Institute of MIT and Harvard, and Massachusetts General Hospital have developed an AI tool called Tyche that can generate multiple plausible segmentations for medical images. Unlike existing models that provide only one answer, Tyche allows users to select the most appropriate segmentation based on their purpose. It does not require retraining and can be used for various tasks like identifying lesions in lung X-rays or anomalies in brain MRIs. The researchers modified a neural network architecture to build Tyche, and testing showed that its predictions were diverse and often better than baseline models. The research was funded by the National Institutes of Health and Quanta Computer.

MIT researchers have developed a new machine learning technique to enhance the red-teaming process, which involves testing AI models for safety. The approach involves using curiosity-driven exploration to encourage the generation of diverse and novel prompts that expose potential weaknesses in AI systems. This method has proven to be more effective than traditional techniques, producing a wider range of toxic responses and improving the robustness of AI safety measures. The researchers aim to enable the red-team model to generate prompts covering a greater variety of topics and explore using a large language model as a toxicity classifier for compliance testing.