Beyond CES: Samsung, Google and Qualcomm are kicking off the race for AI-powered devices

With the new Galaxy S24, Samsung doubles down on AI smartphones

After making big AI bets last week in Las Vegas during CES 2024, Samsung anted up again at yesterday’s Samsung Unpacked event. Along with debuting the Galaxy S24 smartphone, Samsung also announced its first smart ring and a new generative AI model that combines on-device and cloud-based computing. Some of the new Samsung AI features unveiled yesterday included real-time translation and AI-generated text messages, AI-assisted photo and video editing. 

Many of the new features are powered by Google’s latest Gemini AI models that were announced in December. Another new AI feature for the Galaxy S24 will be “Circle to Search,” which lets Android users select something on their screen — by circling, highlighting, scribbling or tapping part of it — without leaving whatever app they’re using. (Samsung also hinted at new AI features last week with a sponsored takeover of The Sphere in Las Vegas that turned the venue’s exosphere into a massive out-of-home ad featuring Marvel’s Doctor Strange.)

“Our phones are really a window to the world’s information,” Cathy Edwards, Google’s VP of Search, said onstage at Samsung Unpacked. “And because humans are naturally curious, you probably come across things in your apps all the time that you want to learn more about. But when you’re immersed in a moment of discovery, it can feel disruptive to stop what you’re doing and switch to another app to learn more.”

New AI tools will likely raise new questions about what the features mean for user privacy. According to Edwards, only the selected part of a screen will be seen by Google when someone uses Circle To Search. Privacy protections were also brought up briefly by Samsung VP of Product Management Drew Blackard. He noted Galaxy’s security and privacy controls dashboard show which apps accessed location data or camera information and allows users to remove access with “just a tap.”

Google expands its own AI galaxy

Many of Google’s AI features will also appear on other devices powered by Android operating system, which also had a large footprint last week at CES. Outside the Las Vegas Convention Center, CES attendees tried Google’s GenAI tools to create physical postcards that Google offered to mail anywhere in the world. Other demos included creating AI-generated phone wallpaper and testing business-focused GenAI tools for laptops. 

This year is expected to deliver big updates for on-device AI features from other tech giants on various laptops and smartphones. Some of the devices that debuted at CES included a new Microsoft Copilot key, which will be built into devices from several brands including Samsung, Dell and Lenovo. The wave of AI devices will enable all sorts of new features for consumers and companies that uses them. 

Despite the flurry of AI product debuts, many aren’t available quite yet. Some of the new AI-enabled PCs won’t be coming out until this summer when devices arrive on the market with the latest Snapdragon chip from Qualcomm. Last week, the semiconductor company gave CES attendees a chance to demo how on-device processing for voice and text GenAI prompts can be faster than only using cloud computing. (Samsung, which will use Qualcomm’s Snapdragon 8 Get 3 process, says its AI features for the Galaxy S24 will roll out in the first half of 2024.)

“There’s no way the cloud world can handle all that in a sustainable, power efficient, economic fashion,” Qualcomm CMO Don McGuire told Digiday in December. “When a ChatGPT prompt costs 7 to 10 times more than a Google search, you know there’s going to be a problem at scale. That’s where on-device really comes in.”

Putting AI behind the wheel — but not yet in the driver’s seat

This year will also see new AI features from major carmakers as several brands integrate new AI-powered voice chatbots. Volkswagen at CES debuted its new ChatGPT-enabled voice AI called IDA, which will arrive in the second quarter in collaboration with the startup Cerence. Meanwhile, Hyundai said it’s working on new AI features with its own large language model; Mercedes-Benz plans to add a new voice AI assistant to its in-car operating system; and Microsoft is working on new AI features with for the new Afeela EV being developed by Sony and Honda. Similarly, BMW announced a new voice AI integration powered by Amazon’s new Alexa LLM. The German brand is also creating new in-car augmented reality features in collaboration with the AR glasses maker Xreal. 

“Everything — all the car functions like navigation, climate, calling, messaging, music — everything that’s controlled today is controlled through this new system,” Cerence CTO Iqbal Arshad told Digiday last week at CES, adding that other LLMs beyond ChatGPT could be added later. “It’s a very extensible system and it’s pretty complete today.”

Beyond smartphones, laptops and cars, other devices will make use of AI and other emerging tech in 2024. New mixed reality headsets and are arriving soon, including Apple’s Vision Pro in February and Xreal’s Air 2 augmented reality glasses in March. Another new device is an AI-powered gadget called Rabbit R1 that’s smaller than a smartphone. After making a splash at CES, Rabbit says its device — which promises to handle a range of actions usually done by other apps — has already been pre-ordered by more than 40,000 people and backordered through mid-summer.

Even grocery carts are getting smarter, which is in some part due to the rise of retail media networks among media choices. Some grocery chains have been testing an AI-equipped shopping cart from Instacart called Caper Cart, which lets people scan, weigh and pay for items, while also providing personalized product ads, seasonal recommendations and guide shoppers around a store. 

“What I’m really excited about on the Instacart marketplace and on the Caper Carts is the AI is all behind the scenes,” said Ali Miller, Instacart’s senior director of product management. “Of course chatbots are exciting, everyone is excited to see what they can do, but I think it’s really powerful when you don’t need to think about a new type of modality.”