Page Content

One way to streamline image description is to use AI tools primed for image descriptions including the ones listed below. No matter which tools you use, there are some tips and cautions listed below.

Caution: Always review image descriptions for accuracy and appropriateness.

Penn State Image Description Tools

The Penn State tools are recommended for preserving security and optimizing copyright protections.

Microsoft Office and Copilot

Microsoft Office and Copilot both provide robust description options through the Penn State Microsoft license. No extra tokens or subscriptions are needed.

Microsoft Office

To access these tools in Microsoft Word, PowerPoint, Outlook, Excel and elsewhere:

  1. Right click any inserted image within a document.
  2. Follow the instructions to insert image alt text in Microsoft Office.
  3. Click the Generate alt text button to use the AI tool.
  4. Audit the text then click the Save/Approve buttons as needed.
    Note: This step is necessary within the Microsoft Accessibility checker, but not the stand alone alt text window.

Copilot Chat

This option allows you to refine description options. It can also provide estimates of data points for simple graphs and full transcriptions of embedded text. To access the options:

  1. Log in to the Penn State Microsoft Office service. You will be directed to the Copilot chat interface.
  2. Click the + sign on the left side of the chat window to upload a file.
  3. You can ask the chat to describe the image and modify parameters as needed.
    Note: A useful starter prompt is "Describe the image to someone who cannot see it."
Note: World Campus has uploaded an "Image Description Agent" available in the Penn State Microsoft Agent Store. This includes preset questions and utilities.

AI Studio (Chat GPT)

AI Studio provides Penn State users access to the Chat GPT engine which is also also has powerful image description and image contrast analysis capabilities.

Note: AI Studio provides a monthly token quota for each user. Do tests of image description in multiple services to determine which is the most cost effective for your needs.

  1. Log on to AI Studio with the Microsoft Authentication to access your account.
  2. In the agent menu on the upper left, select a ChatGPT agent from Open AI.
  3. Select the "hamburger" icon on the upper left to open up an options menu.
  4. Select the option for a New Chat.
    Note: Creating new chats helps an AI engine focus efforts and reduce costs.
  5. To label a chat, input the topic name then "No response needed." This tells the AI not to start a search or drawn an image, saving tokens.
  6. Click the + sign on the left side of the chat window to upload a file.
  7. You can ask the chat to describe the image and modify parameters as needed.
    Note: A useful starter prompt is "Describe the image to someone who cannot see it."

General Tips

  1. Start a new chat for each image or set of related images uploaded. A clear chat can help AI focus on the current task.
  2. Audit descriptions for accuracy and appropriateness. Remove extra interpretations of the image such as an "overall message" or descriptions of people’s emotional state.
  3. Make sure your image file is as clear file as possible. You can upload enlarged images, unrotated images or cropped images to make the description task easier.
  4. You can use AI to transcribe embedded text in an image. This could be a seperate command.
  5. AI usually cannot identify specific people (e.g. celebrities) out of privacy concerns. A tool like Google reverse image search may able to do that task more quickly.
  6. AI may not provide descriptions of images classified as being offensive (e.g. racist propaganda or images with partial nudity). Explaining your task may be able to override that restriction.
  7. AI is accurate with charts, graphs and many maps, but may lack information on niche subject areas.

Top of Page

Last Update: July 1, 2026