Samsung says regional AI restrictions won’t leave Galaxy users dependent on one model

Samsung says it is preparing for a scenario where access to a major AI model such as Google Gemini could change across markets, with the company increasingly relying on a multi-model strategy and regional partnerships to ensure its Galaxy AI experience is not dependent on a single provider.


Speaking during a media roundtable, Jay Kim, Corporate EVP and Head of Customer Experience Office, MX Division, Samsung Electronics, said the company already works with multiple AI models, including its own models, open-weight models and frontier models. “As I said, we have our internal model. We have an open-weight model and we have a frontier model. We're actually working with all of our partners,” Kim said.



He added that Samsung's approach varies depending on the market, with the company also working with local partners in different regions.


“We understand how the AI is impacting globally and how all the markets, the governments are addressing the issues. We are a global company and we manage all the global markets. So, it's not just that we're managing the models focused in the US,” Kim said.


The comments come as AI models are becoming an increasingly important part of the smartphone experience, while governments around the world are also paying closer attention to the development and deployment of foundation models.


For Samsung, the challenge is not simply about ensuring that one particular AI model remains available. The company is looking at how different models can be integrated into its devices depending on the market, regulatory environment and the requirements of individual services.


Kim said Samsung is working with multiple partners globally to balance the needs of different regions and is also engaging with governments on regulatory requirements. “It’s an industry-wide issue,” he said, adding that Samsung is focused more on understanding how AI technology evolves, how models are managed and how they eventually connect with consumers.

Samsung is not relying only on Gemini

Samsung has increasingly positioned Galaxy AI as a combination of its own software and services with AI models from external partners. Kim said the company is already using multiple models on its devices rather than depending exclusively on Gemini.


“We're actually already using multiple models on our devices. It's not just Gemini,” he said. According to Kim, Samsung chooses models depending on the service and entry point, using what it considers the most suitable model for a particular use case.


“The models are evolving…one model comes and then the next model is better and then another one comes. It's more of a race,” he said.


This approach could give Samsung greater flexibility as the AI industry evolves. Instead of tying Galaxy AI to a single foundation model, Samsung can potentially switch between models or combine different models depending on the requirements of a service.



Kim said the focus is ultimately on what the technology means for consumers rather than the model itself.

Samsung wants AI to work without consumers thinking about AI

The company also said its approach to Galaxy AI is increasingly focused on making AI invisible to the consumer.
Kim said Samsung does not necessarily want users to understand whether a particular function is powered by AI. Instead, the objective is to reduce the friction involved in completing everyday tasks.


“We don't need our consumers to understand whether this benefit is coming from AI. They just need to understand that what they want to get done, we just make sure that they get it done,” he said. Samsung has been moving in this direction with features such as Circle to Search, Photo Assist, Creative Studio, Cal Assist and Live Translate.


The company said it measures these features differently depending on their intended use. While frequent usage is important for some features, others may be used less often but become important when consumers actually need them.

India is becoming an important AI market

Samsung also highlighted the growing adoption of Galaxy AI in India. JB Park, CEO and President of Samsung Southwest Asia, said AI usage among Indian consumers is higher than the global average. According to Samsung, features such as Photo Assist, Creative Studio and Audio Eraser are used nearly twice as often in India as they are globally.


The company is also working with IITs, startups and its R&D centres in India to improve AI usability and develop use cases suited to local consumers.


Samsung's R&D centres in Noida and Bengaluru have also contributed to features aimed at Indian users. One example is Voice Focus, which uses AI to detect background noise during calls and automatically activate the relevant mode. Samsung has also added Hindi and Gujarati to Galaxy AI, with the company saying native speakers are using features such as Call Assist and Interpreter.


Park said Samsung is continuing to work on expanding the linguistic reach of Galaxy AI, particularly given India's large number of languages and dialects.

Hybrid AI will remain Samsung's approach

The company does not see the future of AI as being entirely on-device or entirely in the cloud. Kim said Samsung's approach is hybrid, with certain services running locally on the device while more computationally intensive tasks that require deeper reasoning, global information or access to larger datasets can be handled through the cloud.


He said Samsung does not intend to define a fixed percentage of AI processing that will always happen on-device.
Instead, the company expects the balance to depend on the consumer, the task and the hardware available. The approach also becomes important as Samsung expands AI beyond premium smartphones into watches, glasses and other form factors.


Kim said the smartphone will continue to act as a central computing hub, particularly for devices such as AI-powered glasses where processing power and battery capacity are constrained.

The bigger challenge is keeping AI experiences consistent

Samsung acknowledged that maintaining a seamless AI experience across phones, watches, glasses and other devices remains difficult. Kim said the company does not yet have the “right answer” for this challenge, but believes a hybrid model will be important. Galaxy AI is currently positioned for flagship devices, while Samsung is also developing what it calls Awesome Intelligence for its mass-market products.


The company's objective is to gradually cascade capabilities from its flagship devices to more affordable segments while balancing on-device processing, cloud capabilities and hardware limitations.


For Samsung, the AI race is therefore no longer simply about having the latest model. It is increasingly about how quickly and safely those models can be integrated into an ecosystem spanning different devices, markets and regulatory environments.

And in the event that access to a model such as Gemini changes in a particular region, Samsung's bet appears to be that having multiple models, partners and its own AI capabilities will allow it to keep Galaxy AI running without making the consumer deal with the complexity behind it.