Version 34.0

Written & Compiled by Macklin Andrick, GPJ Sr. Creative Technologist

GPJ’s Experience Technology practice is your trusted guide through the latest technological shifts and how they might benefit your next experience. Another week, another opportunity to explore some of the weird and wild things going on in the tech world!

The UN says “I do” to AI ethics with a global resolution, and MrBeast turns game show dreams into a $5M reality check. Meanwhile, the brainy bunch at IBM and NASA are making “science” a less scary word.

World’s first global AI resolution unanimously adopted by the UN

On March 21, 2024, the United Nations General Assembly unanimously adopted a groundbreaking resolution on artificial intelligence (AI), designed to protect personal data, enhance privacy, monitor AI risks, and uphold human rights. Initiated by the United States and supported by China and 121 other countries, this nonbinding agreement—titled “Seizing the opportunities of safe, secure and trustworthy artificial intelligence systems for sustainable development”—marks a significant step toward establishing international norms for AI development. Despite being nonbinding, the resolution has been well received within the AI industry, with key industry figures expressing strong support for its goals. The resolution is part of a broader effort to guide the development of AI technologies in a manner that safeguards democratic processes and human rights, amid growing concerns over AI’s potential negative impacts. This initiative underscores a remarkable level of international cooperation and consensus on the importance of ethical AI development, despite initial resistance from certain nations.

State of Unreal 2024: All the news from Epic’s GDC showcase

During the 2024 Game Developers Conference, Epic Games presented a range of updates at their State of Unreal keynote, including the introduction of Lego templates in Fortnite Creative and the Unreal Editor for Fortnite (UEFN), allowing players to create and monetize their own Lego-themed games within Fortnite’s ecosystem, subject to content ratings. Additionally, Epic unveiled the integration of their hyperrealistic MetaHumans into UEFN, enhancing the tool with new camera options and physics sandbox gameplay features. The company also previewed the mobile version of the Epic Games Store, set to launch on iOS and Android, and debuted the first trailer for “1943: Rise of Hydra,” a narrative-driven action-adventure Marvel game directed by Amy Hennig. These developments signify Epic’s continued expansion and innovation in the gaming and digital content creation spaces, leveraging their technology to offer new creative possibilities and experiences to a broad audience.

IBM and NASA build language models to make scientific knowledge more accessible

In an innovative collaboration, IBM and NASA have developed a suite of efficient language models trained on a comprehensive corpus of scientific literature, leveraging transformer architecture to significantly enhance the accessibility of scientific knowledge. These models excel in a variety of tasks including classification, entity extraction, question-answering, and information retrieval across multiple scientific domains. Achieving notable improvements over existing models in benchmarks and specialized tasks, such as a 5% increase in biomedical task performance and marked enhancements in scientific question-answering and Earth science entity recognition, these models utilize a novel approach for document retrieval through retrieval augmented generation (RAG). They stand out for their ability to generate rich embeddings for effective information retrieval, demonstrating superior performance in retrieving relevant passages from scientific texts. Now available on Hugging Face, these models represent IBM and NASA’s commitment to open and transparent AI development, providing the scientific and academic communities with powerful new tools for leveraging scientific knowledge, underscoring the potential of specialized training and innovative methodologies in advancing the field of AI.

MrBeast and Prime Video announce ‘largest game show in history’

MrBeast, YouTube’s biggest content creator, has announced a collaboration with Amazon Prime Video to create what is being dubbed as the “largest game show in history.” The show will feature over 1,000 contestants vying for a $5,000,000 prize, setting new records in the process. This move to Prime Video marks a significant step for MrBeast, who is known for his extravagant YouTube videos that often involve large sums of money and elaborate challenges. With a subscriber base of 245 million, MrBeast has previously recreated games like “Squid Game,” offering substantial cash prizes, demonstrating his ability to produce content with the scale and budget of traditional TV shows. The announcement underscores the trend of digital creators moving into mainstream media, a transition that has seen varied success in the past. However, MrBeast’s unique position in the cultural landscape, particularly among younger audiences, suggests that “Beast Games” could significantly expand his reach and influence.

Adobe Substance 3D turns text into backgrounds and textures

Adobe has introduced two groundbreaking AI-powered features in its Substance 3D design software suite, aimed at significantly enhancing the workflow of 3D artists by utilizing text descriptions to generate creative assets. At the Game Developers Conference, Adobe unveiled the “Text to Texture” and “Generative Background” tools, leveraging its Firefly AI model. The “Text to Texture” feature, available in the Adobe Substance 3D Sampler 4.4 beta, enables the creation of both photorealistic and stylized textures—like scaled skin or woven fabric—from simple text prompts, which can then be applied to 3D models. The “Generative Background” tool, part of Substance 3D Stager 3.0 beta, uses text prompts to craft background images for staging 3D scenes. Interestingly, these innovations rely on 2D imaging technology to produce images that are applied in a manner to give a 3D effect. Both features, currently free during their beta phase, are developed to streamline the design process, making it easier for designers to generate high-quality textures and backgrounds without the need for extensive searches for reference materials, showcasing Adobe’s commitment to enhancing 3D design with AI.

Introducing Quality Novel View Synthesis and 3D Generation from Single Images

Last week Stability.AI released Stable Video 3D (SV3D), a generative model based on Stable Video Diffusion, advancing the field of 3D technology and delivering greatly improved quality and view-consistency. Based on the foundations of Stable Video Diffusion, SV3D adds new depth to their video generation model with the ability to create and transform multi-view 3D meshes from a single input image. “By adapting our Stable Video Diffusion image-to-video diffusion model with the addition of camera path conditioning, Stable Video 3D is able to generate multi-view videos of an object,” the company said in the announcement. “Additionally, we propose improved 3D optimization leveraging this powerful capability of Stable Video 3D to generate arbitrary orbits around an object.” The new model was trained on a very large set of objects and items, making it perfect for e-commerce, retail and gaming use cases.

More Cool Stuff We Found

Missed a Creative Tech Byte? Find all previous versions here.