Understanding generative AI.

Defining generative AI.

highly detailed little bird on a cobble street with palm trees

Prompt: highly detailed little bird on a cobble street with palm trees

AI vs. generative AI.

Why generative intelligence is so intelligent.

three labradoodle puppies run on the grass

Prompt: three labradoodle puppies run on the grass

Applications of generative AI.

Corporations and generative AI.

Individuals and generative AI.

interior Design, a perspective of of a living room and a kitchen with an island, large windows with natural light, Light colors, vegetation, modern furniture, skylight, modern minimalistic design

Prompt: interior Design, a perspective of of a sitting room and a kitchen with an island, large windows with natural light, Light colours, vegetation, modern furniture, skylight, modern minimalistic design

Limitations and challenges of generative AI.

Generative AI’s capabilities are so astonishing that it can be easy to lose sight of its limitations. Here are several challenges to overcome.

The AI isn’t always right.

As we talked about in the section “Applications of Generative AI,” generative AI tools like ChatGPT are not always factually accurate. There may come a time when finetuned datasets and algorithms reduce the risk, but in the meantime, we humans must be sceptical consumers of what we read. Validate the information by comparing it to a trusted source.

Bias can be anywhere.

Fact-checking is relatively easy. Blocking societal biases, such as those around gender or race, from generative AI results is more difficult. Yet that too is necessary. To prevent societal biases from appearing in generative AI results, the people responsible for the AI must identify and mitigate bias from design to development to deployment and be committed to ongoing oversight.

As users, we can also help root out bias. Say you enter the text prompt “scientist in a lab coat holding a test tube” into an AI art generator. Do the results only show one type of person, no matter how many times you click the “generate” button? You could send a message to the makers of the generator about the blind spot and then refine your text prompt to produce more diverse results.

scientist in a lab coat holding a test tube

Prompt: scientist in a lab coat holding a test tube

Generative AI can use a lot of energy.

Companies developing generative AI tools should also be aware of the energy currently necessary to train and maintain these tools. The industry is waking up to the need to reduce its carbon footprint, but there’s still a long way to go.

Intellectual property rights are an issue.

Professional creators are rightfully concerned about copyright infringement. These concerns are currently being addressed by the courts. Adobe is one example of a company working to assist creators. In addition to developing Firefly’s generative AI responsibly, Adobe is also helping create industry standards through the Content Authenticity Initiative (CAI) and working toward a universal “Do Not Train” tag that lets creators control whether allow AI models can train on their work.

Integrating generative AI into your workflow.

Embrace the future of design with Adobe Firefly’s generative AI.

a Japanese tea garden

Prompt: a Japanese tea garden

Questions? We have answers.

Dream Bigger with Adobe Firefly.

Imagine, experiment and create with generative AI in the Firefly web app. New to Creative Cloud, now available for commercial use.

You May Also Like