Qualcomm to bring generative AI to phones soon.

Qualcomm to bring generative AI to phones soon.

The Rise of Generative AI on Mobile Phones

Generative AI, such as ChatGPT and Midjourney, has captivated the imagination and disrupted industries, but until now, its reach has been mostly confined to desktop computers. However, next year, generative AI will make its way onto premium mobile phones, thanks to Qualcomm’s top-tier chips. This development promises to revolutionize the mobile experience by bringing advanced AI capabilities to our fingertips.

While phones have long utilized AI for tasks like photo touch-ups and autocorrect improvements, generative AI tools will take it to the next level. Qualcomm is integrating generative AI into its upcoming premium chips, set to be unveiled at the annual Qualcomm Summit in Hawaii in late October. Attendees of the summit will have the opportunity to experience firsthand the transformative potential of generative AI on phones.

Ziad Asghar, Qualcomm’s senior vice president of product management, explained to ENBLE why users should be excited about on-device AI. By having access to a user’s data, including driving patterns, restaurant searches, and photos, generative AI on phones can provide customized and more helpful solutions compared to cloud-based alternatives. This personalized experience is the true promise of on-device generative AI and is the aspect that has the Qualcomm team most excited.

Aside from the personalization aspect, there are other advantages to having generative AI on-device. Privacy is a major concern, and with on-device AI, queries and personal data remain on the phone, avoiding the need to relay information to distant servers. On-device AI is also faster than relying on cloud computation, and it works even in areas with no cellular service, making it perfect for airplane travel or remote locations.

Moreover, deploying on-device solutions makes business and efficiency sense for companies like Qualcomm. As machine learning models become more complex, it becomes increasingly expensive to run servers to answer queries. Qualcomm’s solution is to leverage the existing silicon in devices, eliminating the need for additional computational power at a hefty cost. Running AI on your phone becomes effectively free, as you’ve already paid for the computing power upfront.

The possibilities for on-device generative AI are vast. Qualcomm envisions a future where users can seamlessly tweak images, generate new content, and even have full conversations with generative AI-powered virtual assistants. These advancements could be made possible through collaborations with third-party AI providers like Facebook Meta’s Llama 2. The integration of generative AI into mobile apps opens the door to more personalized gaming experiences, social media enhancements, and content creation tools.

Beyond entertainment and personal use, generative AI has the potential to be a game-changer in various domains. In education, AI could gauge students’ comprehension by tracking their expressions and body language and then tailor the material to their individual needs. In navigation, generative AI could assist blind individuals by recognizing their surroundings and providing real-time guidance.

Qualcomm’s upcoming summit will not only showcase generative AI for phones but also for other platforms like PCs and automobiles. The AI-powered personal assistant could help with everyday tasks, such as scheduling meetings or preparing presentations. Asghar likened this future to the iconic smart car from the TV show Knight Rider, where AI becomes instrumental in facilitating our daily lives.

Qualcomm’s on-device generative AI solutions are made possible by the power of their top-tier chips such as the Snapdragon 8 Gen 2 and the upcoming Snapdragon 8 Gen 3. These chips can handle AI models with billions of parameters, making them suitable for mobile searches. Real-world tests will determine how effectively generative AI can be implemented without excessive battery drain. Asghar expressed confidence in Qualcomm’s silicon, citing the MWC 2023 demo device running queries all day without exhausting the battery or overheating.

While Qualcomm is leading the charge, other tech giants like Apple, Google, and Microsoft are also exploring generative AI. For example, Apple has been quietly improving its on-device AI capabilities and is rumored to be developing its own framework for large language models. Meanwhile, Google and Microsoft are integrating generative AI into their productivity platforms. However, Qualcomm’s chip-based generative AI is likely to make the first significant impact on mobile phones.

As we await the Snapdragon Summit, it’s intriguing to contemplate the changes that generative AI will bring to the mobile experience. Qualcomm’s on-device solutions promise a new level of personalization, efficiency, and innovation. Once generative AI is integrated into mobile apps, we can expect customized gaming experiences, enhanced social media interactions, and streamlined content creation. The full potential of generative AI on mobile phones is yet to be unlocked, but we can be certain that it will have a profound impact on our daily lives.