Generative Artificial Intelligence

Generative AI Tools

As we learn to navigate together the fast-evolving landscape of GenAI and to harness its power, it becomes increasingly important to approach the use of dedicated tools with a balanced understanding of their capabilities and limitations. The applications mentioned here are the current tools of choice to either generate texts or create visual art, and as might be expected, this curated list should be regarded as merely a steppingstone into the expansive realm of GenAI. As you invest more time to explore the world of GenAI, you will learn to recognize the unique attributes of the various tools and their remarkable potentials. Therefore, the ethical and mindful use of these tools must stand as an essential guiding principle. 

Generative AI is a rapidly emerging technology that is impacting higher education. To focus diverse expertise on guidance, support and standardization, Provost John Coleman and CIO Mairéad Martin have charged a new Generative AI Center of Expertise (GenAI CoE). The GenAI CoE will support innovation by defining and implementing infrastructure, procurement pathways and best practices.

Critical aspects of responsible usage involve safeguarding privacy and data security. When interacting with generative AI tools, refrain from inputting sensitive or confidential information. These tools, while powerful, cannot guarantee the confidentiality of the information shared. To gain further insights and guidance on using generative AI tools responsibly, we encourage you to refer to the University of Illinois Digital Risk Management Generative AI page. Here, you'll find valuable resources and best practices that will aid you in navigating this dynamic and transformative field while prioritizing security and privacy.

Image Generation Tools

In the ever-evolving world of digital art and image generation, four standout applications have risen to prominence as the preferred tools for creators seeking to craft innovative visual experiences and produce realistic images. These innovative tools, Adobe Firefly, Midjourney, Stable Diffusion, and DALL-E2, are enabling a new era of digital creation, pushing the boundaries of what is possible in the realm of visual art. View the teaser videos below and explore further with us the possibilities offered by these GenAI tools to dive into the world of digital media.

Text Generation Tools

In today's fast-paced digital landscape, language processing and content creation tools have emerged as transformative forces, reshaping the way we communicate and create content. The latest innovative GenAI tools, such as ChatGPT, Google Bard, Microsoft Copilot, and Perplexity, have taken the world by storm, offering unparalleled generative capabilities. We invite you to view the teaser videos below and to explore with us the full potential of these tools.

Audiovisual AI

Michel Bellini, Director of the Center for Innovation in Teaching & Learning at the University of Illinois at Urbana-Champaign, is recreated using generative AI. By uploading 2 minutes of audio/video and giving permission to a company called HeyGen, he can type in any text and have his avatar read it with his likeness and voice.

Check out the video below.


Audio/Visual Generation Tools

Time-based media is undergoing similar innovations to the text and image realms. Machine Learning tools like upscalers (TopazAI) and "airbrushing video" (Adobe After Effects) have been staples for some time in media production. What's new is how LLMs and GenAI is adding immense fidelity at a fraction of the compute time. 

Text-to-Speech (TTS) tools will synthesize a voice to speak any written content. We've had robotic voices, but now we have incredibly realistic stock voices, and even the ability to clone one's voice (see ElevenLabs or the open-source project, Tortoise). Tools with varied utility are bolting on TTS like the transcription and media editor Descript.

Other tools generate video. Avatar-based platforms like HeyGen and Synthesia concentrate on realism of a character in their library or a submission process to represent the user (see example used in iMBA program). Some tools like Pika serve as a text-to-video (and even image-to-video or video-to-video) platform.

Newer to the scene are text-to-music tools like Google's Lyria and Facebook's MusicGen. Several tools simply combine various stock instruments of the same beats per minute, but these newer tools delve deeper and synthesize while retaining continuity.