Seeing Through AI Eyes: Space Image Analysis

What you need

Steps

1. 01

   Choose a NASA photograph

   ~5 minutes

   Download one of the NASA public-domain images listed below. All are free to use with no restrictions. For the biggest "wow" moment try the Webb Deep Field; for the most personal connection try Earthrise. The Artemis II images are so recent the AI may not have been trained on them — which makes them especially interesting to test.

   What to expect

   A saved image file ready to upload. NASA images are among the most information-dense photographs ever taken — each one contains layers of science, history, and detail that AI can surface.

   Choose an image

   

   Earthrise — Apollo 8, 1968

   

   Buzz Aldrin's Visor — Apollo 11, 1969

   

   Webb's First Deep Field — SMACS 0723

   

   Mars Curiosity — Gediz Vallis Channel
2. 02

   The open-ended question

   ~5 minutes

   Open Claude (or any model). Drag your image into the chat or use the attachment/upload button. Then type the prompt below. Most beginners are surprised by how much the AI sees — it doesn't just say "space photo", it identifies specific structures, explains processes, and connects the image to broader scientific context.

   PromptCopy

   What do you see in this image? Describe everything you can identify — what it shows, what the science is, and anything that might not be obvious to a casual viewer.

   What to expect

   The AI will identify the image if it's a well-known NASA photo, describe visible features in detail, explain the underlying science, and point out things you might have missed — like how colors represent specific wavelengths of light or how scale compares to something familiar.

   Try this

   If the AI misidentifies the image, correct it: "That's actually [image name]. With that in mind, what can you tell me?"
3. 03

   Go deeper with follow-up questions

   ~10 minutes

   The AI remembers the image — no need to re-upload. Ask the follow-up prompts below one at a time. Each one pulls a different layer of understanding from the same image. The child-friendly explanation is especially revealing: watch how the AI simplifies without dumbing down.

   PromptCopy

   What are the specific colors in this image telling us? Are they real colors or have they been enhanced? What wavelengths of light are we actually seeing?

   What to expect

   Each follow-up question surfaces a new dimension: colour science, physical sensation, instrument technology, scientific impact, or accessible explanation. The AI adapts its depth and language to what you ask.

   Try this

   Other follow-ups to try: "If I were standing at the location where this image was taken, what would I actually see with my naked eyes?" / "What was the technology that captured this image?" / "Explain this image to a curious 10-year-old. Make it exciting."
4. 04

   Same image, different eyes

   ~10 minutes

   Upload the same image to a second AI model (ChatGPT, Gemini, or Grok) and ask the same opening question. Compare the two responses: did they identify the same features, did one notice something the other missed, which gave more scientific depth? This comparison teaches you that AI vision models have different strengths.

   PromptCopy

   What do you see in this image? Describe everything you can identify — what it shows, what the science is, and anything that might not be obvious to a casual viewer.

   What to expect

   The two responses will differ. One may be stronger on technical detail, another on narrative explanation, another on connecting the image to broader context. Neither is simply "better" — they have different personalities.

   Try this

   Try a third model for a three-way comparison. Note which model you'd choose for scientific accuracy versus which you'd choose to explain something to a non-expert.
5. 05

   Push the boundaries

   ~5 minutes

   Try something the AI wasn't specifically designed for — but can often do anyway. Pick one of the prompts below. This step shows that AI vision isn't just a label-maker: it can reason about what it sees, make creative leaps, and connect visual information to broader knowledge.

   PromptCopy

   If this image were a painting in a museum, what would the placard say? Write a museum description that captures both the science and the emotional impact.

   What to expect

   A creative, thoughtful response that treats the photograph as an artifact of human achievement. The AI draws on both visual analysis and broader cultural/historical context to write something genuinely compelling.

   Try this

   Alternative prompts: "Based on what you can see, estimate the physical scale — compare it to something on Earth." / "What processes visible in this image also happen on Earth?"

What you learned

• How to upload images to AI models and start a visual conversation
• How to ask progressively deeper questions about a single image
• How different AI models interpret the same visual information differently
• That AI vision goes far beyond identification — it can reason, explain, compare, and create
• A method you can now apply to any image: documents, charts, screenshots, art, nature, anything

Go further

Troubleshooting