By: TM Roh, President and CEO of Samsung Electronics
New technologies follow a familiar arc. They begin as marvels—expensive, experimental and full of fanfare. As the excitement settles, the technologies that truly change society fade into the background. That isn’t because they become less powerful but because they become dependable, affordable and broadly usable—quietly triumphant and easily overlooked. As Amara’s Law says, we overestimate a technology’s short-term impact and underestimate its long-term effects.
Today, artificial intelligence is the great newcomer, as loudly proclaimed as the mobile phone or the internet was. Being mindful of Amara’s Law in its case means advancing AI while never losing sight of what makes it scalable as a practical, everyday innovation. That’s what distinguishes mere invention from infrastructure.
Infrastructure is something people count on to work—reliably, universally and without the need for outside expertise. The core question isn’t awareness of AI. Recent research shows that 86% of mobile users already use it. What matters is how practical and helpful AI is in real life. Does it understand users’ context and intent well enough to earn their trust?
Language was an early proving ground. Translation isn’t useful if it works well in a handful of standard global languages but fails among dialects, accents and real-world contexts. That’s precisely where misunderstandings have real-life consequences.
Accessibility is equally important. Features like real-time captions, image descriptions and simplified summaries are more than nice to have. They’re foundational to understanding and action while delivering what infrastructure is built for: consistency.
Another key measure is trust. AI exists in the most personal areas of our lives—among our messages, photos, documents, finances and even health. People rightly ask if giving up control is necessary to take advantage of AI’s benefits. If it feels like the answer is yes, adoption slows. This isn’t fear of technology, it’s a rational response to risk.
AI infrastructure comes with responsibilities for its builders. Infrastructure must be designed to work reliably in real-world conditions and serve everyone, including the most vulnerable users. By that standard, engineering and design should give priority to reach, openness and confidence.
A clear measure of meaningful innovation is how many people use the technology every day. To become a default experience for the majority, AI has to reach more devices and more people with a consistent high-quality experience.
For AI to be universal, it must work equally well across languages, cultures and contexts, delivering the same level of accuracy, fluency and cultural nuance no matter who is using it or where. But openness goes beyond languages. For AI to be used comfortably by more people, it must be intuitive enough that it doesn’t require learning on the user’s part. People shouldn’t have to think about how to operate AI to benefit from it. That is why the best AI often stays in the background. The less visible AI becomes, the more universal the experience feels.
AI gains confidence by delivering dependable performance that is consistently fast and responsive and by making clear that privacy and security are built in. Secure data and transparent information control, grounded in user choice, shouldn’t be optional. They are essential for AI to function as true infrastructure.
The real challenge for the industry isn’t improving AI literacy; it’s designing AI that people can use without having to make a project of it. The idea of AI as infrastructure must also shape the coming phase of agentic AI, which will carry tasks through to completion, not merely provide answers. Done well, this shift will further reduce friction by handling routine actions and organizing what matters without requiring technical commands or constant intervention. The real value for AI won’t be found in benchmarks or model comparisons. It will show up in everyday moments, when more people can understand, participate and move through the world with ease.
This message contains information which may be confidential and privileged. Unless you are the intended recipient (or authorized to receive this message for the intended recipient), you may not use, copy, disseminate or disclose to anyone the message or any information contained in the message. If you have received the message in error, please advise the sender by reply e-mail, and delete the message. Thank you very much.
