Sponsored by

Hey TEDchats Fam!

Happy Friday! Remember when video games were all about smashing buttons and scoring points? Well, AI's latest move is like that friend who jumps in to play, not to outscore you, but to embark on an adventure together, learning and laughing along the way. It's time to meet your new digital sidekick!

👀 SNEAK PEEK

🚨FEATURE STORY

Google Redefines AI in Gaming

Imagine a video game where the goal isn’t to win, but to play alongside you, learn from you, and even take directions. This isn't a glimpse into a distant future but the reality of Google DeepMind's latest AI innovation, SIMA.

How SIMA is Changing the Game

The traditional AI in video games has often been criticized for being either too predictable or unnervingly perfect, breaking the immersive experience for players. Unveiled by Google DeepMind, the inception of SIMA (Scalable Instructable Multiworld Agent) was rooted in the challenge of creating an AI that behaves more like a human companion rather than an unbeatable adversary or a scripted NPC (Non-Player Character).

Sima represents a paradigm shift in how artificial intelligence interacts within the virtual worlds of video games. Developed in collaboration with eight game studios such as Coffee Stain and Hello Games, SIMA is designed to play games ranging from the open-world madness of "Goat Simulator 3" to the expansive universes of "No Man’s Sky". It operates based on simple human text instructions, showcasing an uncanny ability to understand and navigate 3D environments, essentially serving as a genuine gaming companion — albeit in its infancy.

Teaching an AI to Think

A significant breakthrough occurred when the SIMA team adopted a novel approach to AI training. They simulated environments in Unity-based settings where the AI had to interpret human instructions and predict actions. Subsequently, exposing it to real gameplay scenarios with human players led to SIMA demonstrating an impressive ability to adapt and learn. You can read the precise steps taken by Google to develop SIMA in their technical report, available here.

Now, SIMA operates with just two inputs: the visual information displayed on the screen and simple, natural-language instructions provided by the gamer. It uses keyboard and mouse outputs to direct the central character in the games, carrying out these instructions. SIMA is capable of performing approximately 600 basic tasks, each typically requiring about 10 seconds to complete. These tasks encompass navigation commands such as 'turn left', interactions with objects like 'climb the ladder', 'drive the car', or 'jump the fence', as well as menu actions like 'open the map'. Video demonstrations of SIMA can be found here.

Unlocking New Potentials for Gamers and Humanity?

While SIMA is still in its early stages, its potential to revolutionize gaming experiences with a new form of AI companionship is significant. SIMA is currently learning to play games and apply acquired skills to new ones, but it cannot speak yet, unlike advanced AI NPCs. Google anticipates that as AI technology progresses, SIMA could evolve into a more capable gaming companion, assisting players in achieving victory in increasingly complex tasks.

The journey of SIMA has just begun, and Google sees its potential extending beyond gaming. They express hope that “SIMA and other agent research can use video games as sandboxes to better understand how AI systems may become more helpful”. The implications of using video games to enhance AI's capabilities in understanding and executing complex, context-driven tasks in dynamic environments are profound. From virtual training environments to interactive storytelling, SIMA could indeed serve as the blueprint for the next generation of AI (see Google’s announcement).

🧑‍💻 DAILY PROMPT

🍕 QUICK BITES

Backlash Against Under Armour's AI Commercial

The advertising world is abuzz as Under Armour's recent commercial, featuring boxer Anthony Joshua, draws criticism for claiming to be the "first AI-powered sports commercial." Director Wes Walker proudly showcased the commercial on his Instagram, noting that it was created in just three weeks from concept ideation to delivery. However, industry professionals have pointed out that the ad repurposes content from previous works without crediting the original creators — a list of complaints can be found here. This incident highlights ongoing ethical debates about the integration of AI technology in creative fields and emphasizes the importance of acknowledging artistic contributions (view Under Armour’s AI commercial).

Amazon Introduces AI-Powered Tool for URL-Based Listing Creation

In the coming weeks, Amazon will introduce a new feature allowing sellers to create product listings on its platform by submitting a URL from another website. This tool, powered by generative AI, extracts item details from the provided URL to automatically generate an Amazon listing. Initially targeted at U.S. sellers, this development is part of Amazon's initiative to streamline the listing process, enhancing efficiency by minimizing the necessity for manual entry of product information (see Amazon’s announcement).

High-Tech Measures to Combat U.S. Fentanyl Trafficking

In a notable effort to combat U.S. drug smuggling, U.S. Customs and Border Protection (CBP) is poised to deploy advanced AI-powered X-ray machines for scanning vehicles entering the country, targeting fentanyl and other narcotics, as reported by The Verge. While specifics remain unclear, this initiative appears to be part of a broader security overhaul supported by President Biden. In his State of the Union address, Biden articulated plans to add "100 more high-tech drug detection machines to significantly increase the ability to screen and stop vehicles from smuggling fentanyl into America". With over 43,000 pounds of fentanyl intercepted last fiscal year, the U.S Department of Homeland Security (DHS) has released its strategies for combating drug trafficking. According to the report, machine learning models have already led to 240 seizures, encompassing thousands of pounds of cocaine, heroin, methamphetamine, and fentanyl (read the DHS's report here).

🧰 AI TOOLBOX 

🤯 MIND SPARK

ChatGPT Gets a Body! OpenAI and Figure have collaborated to provide ChatGPT with a physical form. OpenAI offers high-level visual and language intelligence, while Figure's neural networks enable rapid, precise, and dexterous robot actions. The results are astonishing — we highly recommend watching the linked video (see Figure’s post).

AI in Medicine: The Yale School of Medicine hosted their first ‘AI in Medicine' symposium last month, featuring an interesting array of panel discussions, rapid talks, and presentations. Video summaries of the event have been consistently uploaded over the past weeks, offering comprehensive coverage now available for viewing (watch Yale's ‘AI in Medicine’ videos here).

The EU Finalizes Rules for the Artificial Intelligence Act! Bans will be imposed on AI applications that threaten citizens' rights, such as indiscriminate scraping of facial images for facial recognition databases, emotion recognition in workplaces and schools, social scoring, predictive policing, among others. However, there are still hurdles to overcome before they take effect, with implementation occurring 6-36 months afterward (see EU Parliament's announcement).

Hire a world class AI team for 80% less

Trusted by leading startups and Fortune 500 companies.

Building an AI product is hard. Engineers who understand AI are expensive and hard to find. And there's no way of telling who's legit and who's not.

That's why companies around the world trust AE Studio. We help you craft and implement the optimal AI solution for your business with our team of world class AI experts from Harvard, Princeton, and Stanford.

Our development, design, and data science teams work closely with founders and executives to create custom software and AI solutions that get the job done for a fraction of the cost.

P.S. Wondering how OpenAI DevDay impacts your business? Let’s talk!

🎨 AI ART