Generate Variations - Trust and Privacy in Edge Delivery Services

Transcript

AI Functionality in AEM Edge Delivery Services Document authoring allows for both copy and image generation. For copy generation, Adobe is using Microsoft Azure, OpenAI, and GPT-4, so GPT is used for the underlying model, which allows for copy generation. For image generation, Adobe is using Adobe Firefly, which is our proprietary image generation technology, and the way Firefly is actually incorporated, is that Adobe has an integration into Adobe Express inside of the tool, and if you want to generate an associated image, you can simply click a button. It then opens up a window for Adobe Express, where you can then use text image to create the rich images as well. Very cool! Another important question, how is Adobe ensuring trust and privacy for the customer data? Is the data being used to train the model? Here, none of the customer data is being used, and one big objective of our customers is to use their brand-specific guidelines, their inputs, to really get the right output that they need. Of course, these customers also do not want to use their data or give their data to the models. They also want to make sure that any proprietary data does not go somewhere else. Adobe stores all customer data in isolation. Everything’s done at the prompt level, and nothing is actually going and getting stored back into the model. This means that the customer data is isolated and protected. Adobe has an agreement with Microsoft Azure to make sure that all the trust and privacy is at the forefront. Because of this, customers should have no concerns about using this feature, and using their data will provide the absolute best output they need.