This blog post was originally published at Qualcomm’s website. It is reprinted here with the permission of Qualcomm.
Generative AI has unlocked the power of AI, and the potential is unlimited
Since the introduction of ChatGPT in November 2022, the world has become enamored with the potential for artificial intelligence (AI). ChatGPT ushered in the era of generative AI, the ability to create new content based on simple text or voice prompts utilizing models trained on vast amounts of data.
Until the introduction of large language models (LLMs), traditional AI applications and solutions analyzed data and made predictions based on learning from a data set limited in size. Traditional AI is already in use across industries for applications like autonomous system control, chatbots, customer service agents, fraud detection, game playing, natural language processing, recommender systems, risk analysis, system management and security surveillance. The applications for traditional AI will continue to grow, but generative AI opens an entirely new world.
With the power to learn from vast amounts of data, LLMs have the ability to recognize, summarize, predict, and generate novel text, audio, images and even video. This brings AI into the realm of capabilities previously limited to humans.
Content creation
The initial push for generative AI is in content creation: the ability to modify existing content and/or create unique, new content. In text form, this could take the form of something as simple as an email or something as complex as an article, blog post, poem or story.
In audio format, a generative AI response could take the form of a voice response, a series of sounds or even a complete song. In image format, this could be a realistic or fanciful picture, graphic design, and creating or removing elements of a picture. These capabilities can also be extended to a video format, which is just a series of images, and combined with audio, the content could take the form a game, movie, a synthetic game character or even a complete metaverse environment.
Educational assets can become less cumbersome to create with generative AI as well. For example, healthcare professionals can use generative AI for patient rehabilitation training, and educators can use generative AI for the development of optimized learning plans.
Design is an area that holds huge potential for generative AI. Generative AI welcomes the ability to combine information about materials, style, environmental information, and engineering knowledge to design clothes, cars, consumer products, buildings and more. This offers the potential for unique or customized designs, reduced time to market, more flexibility and increased reliability to any enterprise that designs a product of any size or shape, while adding a new tool for do-it-yourself (DIY) projects.
The potential for AI content creation is almost endless.
Personalization
Generative AI will have a significant impact on society through personalized digital assistants that have contextual awareness and the ability to communicate in more human-like manners.
To use an example from a recent interview with Durga Malladi, SVP & GM, Technology Planning & Edge Solutions, Qualcomm Technologies, Inc., imagine this assistant, accessible from any number of devices like your car or smartphone, being able to recommend a restaurant based on your personal preference, dietary restrictions and location; make a reservation for you at your likely arrival time; and possibly even place your dinner and drink order.
Generative AI could be used to develop your diet, exercise routine, entertainment calendar and even pay your bills.
Code generation
A key application of generative AI is the creation of code, the lifeblood of all the technology that runs the world around us.
Programmers assisted by generative AI have the potential to create, test, update and maintain code over the life of an application, eliminating millions of man hours and ensuring timely updates.
This is becoming increasingly important as code complexity and size increase with the use of autonomous machines and increasingly sophisticated applications.
Natural user interface
Generative AI is bringing an unprecedented change in how we interact with devices. Natural language understanding through text or speech has dramatically improved, allowing for natural chat-like interactions with devices to accomplish tasks.
Voice as opposed to touch can be the more intuitive interface to interact with devices. Being able to chat with a device rather than figuring out which app is where, how to use it or how to find a setting, is much simpler. It hides the complexity and democratizes the experience for users.
Generative AI can speak multiple languages, not only English, Chinese or Spanish but also programming languages like binary code, images or just about any modality you can think of. This means that multiple apps, programs and devices can work seamlessly together to solve the interface challenge for machine-to-machine communication — and improve your life.
So, it can be much easier to check your car’s tire pressure, decipher the error message on the washing machine, or limit the blue light on your phone at night, thanks to generative AI.
Research
Generative AI has almost unlimited potential in scientific research and analysis. The ability to recommend, model and simulate interactions holds huge potential for drug research, chemical analysis, material design and even atmospheric modeling.
Generative AI could even be used to create simulated data to further improve research and analysis when existing data is limited or difficult to acquire. As the use of generative AI continues, so will scientific breakthroughs.
The possibilities are endless
As with any new and revolutionary technology, the impact across applications, usage models and industries will reach farther than what anyone predicted.
The number of applications for generative AI at this point is difficult to fathom.
Some will even include applications currently using traditional AI models. All of this is possible today and will only improve and expand as the models get refined and optimized. These capabilities will be available in ever increasing personalized versions and on various platforms.
Whether in the palm of your hand, on your laptop or in the car, Qualcomm Technologies is working with the industry to ensure that the power of on-device generative AI can be harnessed in a secure, private, socially responsible and immediate way.
Pat Lawlor
Director, Technical Marketing, Qualcomm Technologies, Inc.
Jerry Chang
Senior Manager, Marketing, Qualcomm Technologies
