📌 1. Thaler v. Perlmutter (U.S., 2023) – DABUS AI Case

Facts:
Stephen Thaler attempted to register copyrights for works autonomously created by his AI system, DABUS.

Holding:

Only human authors can hold copyright in the U.S.

Works generated solely by AI without substantial human input are ineligible for copyright protection.

Implications for E-Commerce Visuals:

AI-generated product images or promotional banners that are fully automated may not qualify for copyright.

Human designers or marketers must make significant creative choices—composition, color schemes, branding style, and layout—to claim copyright.

Key Principle:

AI is a tool; human creative control determines copyright eligibility.

📌 2. Zarya of the Dawn / Midjourney Registration Denial (U.S., 2023)

Facts:
The U.S. Copyright Office denied registration for artwork produced primarily by AI with minimal human prompting.

Holding:

Minimal prompts do not constitute sufficient human authorship for copyright.

Implications:

Simple “generate product image” commands to AI without human curation are unlikely to create protectable e-commerce visuals.

Editors or designers must intervene substantially, e.g., by adjusting lighting, angles, product arrangement, or background settings.

📌 3. Andersen, McKernan & Ortiz v. AI Art Platforms (U.S., 2025)

Facts:
Artists sued AI art platforms for using copyrighted material without authorization to train AI models.

Holding:

AI systems trained on copyrighted works without consent can be liable for infringement.

Implications for E-Commerce Visuals:

AI generating product images using training datasets that contain copyrighted photography, designs, or graphics may create derivative works and legal exposure.

Using licensed or public-domain datasets is crucial to minimize risk.

📌 4. Authors Guild v. Google Books (U.S., 2015) – Transformative Use Doctrine

Facts:
Google scanned books for indexing, which was ruled as fair use due to the transformative purpose.

Implications:

AI-generated visuals for e-commerce may qualify as transformative if they significantly alter or remix source content for creative marketing purposes.

Simply replicating copyrighted product photography or illustrations without modification is riskier.

📌 5. Feist Publications v. Rural Telephone Service (U.S., 1991)

Facts:
The U.S. Supreme Court held that facts themselves are not copyrightable, only original expression.

Implications:

AI-generated visuals representing factual product specifications (e.g., dimensions, color swatches) are less likely to infringe copyright.

Creative expression in photography, stylized renderings, or branded promotional visuals is protected and requires proper authorization if based on existing copyrighted material.

📌 6. Naruto v. Slater / Monkey Selfie Principle (U.S., 2018)

Facts:
A monkey’s photograph could not be copyrighted because it lacked human authorship.

Implications:

AI-generated e-commerce images without human intervention are similarly not eligible for copyright.

Ownership attaches to the human editor, designer, or marketer directing the AI.

📌 7. Klinger v. Google (Australia, 2018)

Facts:
The court addressed whether automated text/data mining infringed copyright.

Holding:

Automated reproduction of copyrighted material can infringe if it reproduces substantial expressive content.

Implications:

Using AI to replicate copyrighted product imagery or promotional designs for e-commerce can create liability if the output reproduces substantial elements of the original works.

Human modification and transformation of AI outputs reduce infringement risk.

🔍 Key Legal Principles for AI-Generated E-Commerce Visuals

1. Human Authorship is Critical

AI alone cannot hold copyright.

Significant human intervention in design, composition, and branding decisions is necessary.

2. Derivative Works

AI trained on copyrighted images may produce derivative works, potentially infringing copyright.

Licensed or public-domain training sets reduce exposure.

3. Transformative Use

AI-generated visuals that add originality, style, or commentary may qualify as transformative.

Direct copying or minimal modification is more likely to infringe.

4. Factual Content vs Expression

Visuals conveying factual product information are less likely to be protected.

Creative elements like photography style, rendering effects, or branding elements are protected.

5. Documentation

Keeping records of human creative decisions—prompt adjustments, editing choices, layout design—strengthens legal claims and copyright ownership.

🛠 Practical Applications in E-Commerce

Scenario 1: Human guides AI to create a stylized product catalog → copyright likely.

Scenario 2: AI generates product images automatically from existing images without human editing → copyright weak or absent.

Scenario 3: AI trained on licensed image datasets → safer for commercial use.

Scenario 4: AI replicates competitor’s copyrighted images → high infringement risk.

✅ Conclusion

AI-generated e-commerce visuals exist on a human-AI collaboration spectrum.

Copyright protection depends on substantial human creative input, transformative use, and careful licensing of training materials.

Commercial use without human authorship or proper licensing exposes businesses to infringement claims.