This Week in AI: The Era of Visual Intelligence

As part of its annual iPhone upgrades, Apple introduced a new feature for its iPhone 16 line to take on Google Lens, an app whose tagline is, "Search what you see." Apple's feature, called Visual Intelligence, is designed to provide relevant information based on where you point your phone's camera.

Users activate Visual Intelligence by click-and-holding the new Camera Control button on the iPhone 16 line, which is placed below the power button on the device's right side. If pointed at a flyer for an event, Apple says its AI will automatically add information to your calendar. It can even identify a dog breed or help with your homework, Apple said. 

"We wanted to give you the ability to instantly learn about everything you see," said Craig Federighi, Apple's software head, during the company's launch event Monday.

But comedian Stephen Colbert was particularly impressed by Apple's example that when pointing the iPhone's camera at a restaurant on the street, the screen may fill with details such as ratings, menu items and hours. 

"Groundbreaking!" Colbert declared during The Late Show on CBS. "Finally, you can stand outside a restaurant and find out when it opens."

AMD's AI push bigger than gaming

The rise of AI and increased attention on GPUs that make it possible, has left some gamers worried that chip companies will stop prioritizing chips and cards specially-designed for home users, and instead focus on businesses, from large customers like Microsoft, Google and OpenAI to ambitious start-ups. 

In interviews with Tom's Hardware, AMD Senior Vice President and General Manager of The Computing and Graphics Group Jack Huynh said the chipmaker is increasingly focusing its efforts on integrating GPUs into its data center offerings. "My number one priority right now is to build scale," he said, adding that Microsoft has praised how well ChatGPT 4 runs on AMD's MI300-series chips.

Despite this AI push, Huynh said AMD still understands the importance of graphics chips for home users. "Don't worry, I love gaming," he added. "We will have a great strategy for the enthusiasts on the PC side, but we just haven’t disclosed it."

Google podcasts based on your notes?

For the first time in a while, people aren't criticizing Google's AI for recommending they add glue to pizza to keep the cheese from falling off. That's because Google unveiled a new service called Audio Overview, which is designed to create "engaging discussions" based on your documents, slides or charts. 

The tool is connected to the company's experimental NotebookLM service, which is designed to allow people to use a chatbot to help sort through their notes. Audio Overview takes the idea that much further, effectively creating podcasts with two "AI hosts" holding a "lively" discussion based on your documents.

"They summarize your material, make connections between topics, and banter back and forth," Google said in an announcement of the feature. "You can even download the conversation and take it on the go."

The service only speaks English for now, Google said, and it may "sometimes introduce inaccuracies." Hopefully it'll at least get my family's pizza recipe right this time.

Adobe previews text-to-video

OpenAI's Sora text-to-video technology is facing increased competition from Adobe and its Firefly AI-image technology. In a new video teaser video published to YouTube, Adobe showed examples of Firefly's text-to-video feature powering short computer-generated shots of animals, nature scenes and fictional characters.

Adobe says its technology can go a step further though, helping extend clips of non-AI videos in its Premiere editing app, or creating a video from a photo uploaded to its service.

Adobe said its tools are designed to "help creatives overcome common editing challenges," and promised that its videos are "commercially safe," a not-so-subtle dig at other generative AI tools that have been criticized for building their technology using copyrighted material without permission from authors, artists or creators.

Roblox expands game worlds with AI

In open source AI news, Roblox has announced an AI tool for generating 3D worlds from text. 

The tool is designed to "make it easy to build new game environments on the platform," MIT Technology Review reported. "Even if you don't have any design skills."

Roblox said developers will be able to write something like, "Generate a race track in the desert," and the AI will create one in the company's game world. "Users will also be able to modify scenes or expand their scope—say, to change a daytime scene to night or switch the desert for a forest," the publication added.

While the technology may make game worlds easier to create, developers will still need to actually program the games and experiences that take place inside them. Still, it's a fun use for AI, and on one of the most popular platforms in the world. 

Read more: AI Tools and Tips 

• Getting started with LM Studio: A Beginner's Guide
• How to Use Generative AI in Photoshop
• How to Use AI to Grow Your Small Business
• Meet Claude, the Best AI You've Never Heard of
• What is TOPS? The AI Performance Metric Explained

Ian Sherr is a widely published journalist who's covered nearly every major tech company from Apple to Netflix, Facebook, Google, Microsoft, and more for CBS News, The Wall Street Journal, Reuters, and CNET. Aside from writing, he tinkers with tech at home, is a longtime fencer -- the kind with swords -- and began woodworking during the pandemic.