A quirky new mobile operating system, claiming to tease the future of AI on smartphones, popped up at Mobile World Congress in Barcelona this week.
South Korea-based startup Newnal showed off mobile software that uses both your past and current data to create a personalized AI assistant. But here’s the twist: The AI intends to be… you. It looks like you, acts like you, learns from you, even grows old like you.
The blockchain-based operating system accesses apps and personal accounts, such as Instagram and Gmail, to create a personalized knowledge graph, essentially learning everything about you. This is then used to create a personalized AI avatar that can respond and essentially act just like you would, according to the company.
Unlike other popular AI models such as ChatGPT or Google Gemini, Newnal said in a press release that its users own and train the Web3 AI Newnal OS on their data instead of what’s on the internet, so no two user experiences are the same. The operating system is blockchain-based and encrypted, open-sourced and can be embedded on any device, the company said.
The model “flips the script on less rigorous privacy models by putting the individual user in full, flexible control of their data,” company told CNET in an email.
To demo the technology at MWC 2025 — which also featured foldable devices, ultra-slim phones and camera concepts — the company added the technology to its own proprietary mobile devices. As seen in a video that the tech site Android Authority shared on X, the top of the display features the AI avatar at the top, with the bottom reserved for traditional smartphone tasks, like social media, shopping and email.
Watch this: Samsung’s Stretchable, Rollable and Foldable Screens Are on Display at MWC 2025
Although details are still slim and Newnal’s concept appears to be far reaching, it could hint at what’s ahead for AI on mobile devices.
“They are gimmicks to the extent that they amaze in the capabilities demonstrated within a highly controlled environment, but when you take them on any kind of robust road test, they implode,” said Dipanjan Chatterjee, an analyst at Forrester Researach, drawing ties to the Rabbit devices and the now defunct Humane AI pin.
“But they are also not gimmicks in that … we are now looking at AI technologies that can see, hear and speak,” he added. “What this means is that the future of our primary devices will likely be liberated from the traditional touch screens we are so familiar with on our mobile phones.”
It’s also building on the buzzy idea of agentic AI, where the artificial intelligence model can act on behalf of the user.
Although the emerging concept has vast potential, there have already been some real-world missteps. Still, Chatterjee believes taking risks like this could prove monumental.
“In the grand failures of devices like these is a whiff of what success will look like in the future – think the Apple Newton evolving into the iPhone and iPad experience,” he said. “Technologies like these are like binoculars that give you a glimpse into the future, but the present is still quite hazy.”
