Module 6: Ethics, Copyright & Future Directions

Generative AI has expanded what individuals and organizations can create, but faster production also introduces questions about copyright, ownership, originality, disclosure, privacy, bias, authenticity, and accountability. Responsible content creation requires more than asking whether an AI tool can produce something. Creators also need to consider whether they have the right to use the source material, whether the output is too similar to protected content, whether their own contribution is copyrightable, and whether the audience should be told how AI was used.

This module focuses on practical decisions content creators can make before, during, and after using generative AI. You will examine current U.S. copyright guidance, compare prompts that directly reference protected creative properties with prompts designed to produce more original results, evaluate disclosure and provenance practices, and consider where AI-assisted content workflows are heading next.

Important legal note: Copyright, trademark, publicity, privacy, and related laws vary by jurisdiction and by the facts of a particular use. The material in this module is educational and is not legal advice. When a project carries significant legal or commercial risk, consult the appropriate institutional policy or qualified legal counsel.

Module Learning Objectives (MLOs)

By the end of this module, you will be able to:

MLO 6.1: Explain the role of human authorship in the copyrightability of AI-assisted content.
MLO 6.2: Distinguish copyrightability, infringement risk, licensing, and disclosure as separate questions in an AI content workflow.
MLO 6.3: Revise image-generation prompts to reduce dependence on protected characters, branded properties, and distinctive creative references.
MLO 6.4: Apply responsible practices involving disclosure, provenance, privacy, bias, accessibility, and human review.
MLO 6.5: Evaluate current AI capabilities and emerging directions without assuming that automation replaces human judgment or accountability.

How to use this lecture: Select a section below to expand it. The image examples are based on actual prompts used during the course-development process. Compare the prompt wording with the resulting visuals and focus on how changing the reference changes the level of originality and risk.

Module 6 Topics

Human Authorship • Copyrightability • Infringement • Licensing • Characters & Branded Properties • Prompt Revision • Disclosure • Provenance • C2PA • Bias • Privacy • Authenticity • Accessibility • Responsible AI • Multimodal AI • Connected Workflows • Human Accountability

1. Copyright, Human Authorship & Ownership

Copyright questions involving generative AI are easier to understand when several different issues are separated. A creator can have permission to use a tool but still create an output that raises concerns about another person's protected work. A creator can also make a lawful AI-assisted work without automatically owning copyright in every AI-generated element.

Four Different Questions

Questions to separate when evaluating AI-assisted content
Question	What It Addresses
Can I use the source or reference?	Permission, licensing, public-domain status, copyright exceptions, or other rights affecting the source material.
Is the output too similar to protected content?	Potential infringement or other intellectual-property concerns involving the generated result.
Can I claim copyright in the final work?	Whether the work contains sufficient human-authored expression that is eligible for copyright protection.
Do I need to disclose AI use?	Platform rules, institutional policies, professional standards, contractual obligations, applicable law, and audience expectations.

Current U.S. Copyright Guidance

The U.S. Copyright Office continues to require human authorship for copyright protection. AI assistance does not automatically prevent a work from receiving copyright protection, but the protectable expression must come from a human author.

Human contribution may include:

Human-authored text, artwork, photography, music, or other expressive material incorporated into the final work
Creative modifications made to AI-generated material
Creative selection, arrangement, or coordination of material
Other human-authored expressive choices that are perceptible in the final work

Important correction: Detailed prompting alone should not be treated as a guarantee of copyright ownership. Current U.S. Copyright Office guidance states that the mere provision of prompts generally does not provide sufficient human control over the expressive elements produced by the system.

Example: AI-Assisted Writing

If ChatGPT produces an initial blog draft and the author substantially rewrites the structure and language, adds original analysis and examples, selects and arranges material creatively, and completes the final editorial work, the human-authored portions and qualifying human creative contributions may be protected. The AI-generated portions do not automatically become copyrightable simply because the author selected or prompted them.

Watch & Connect: What Can and Cannot Be Copyrighted With AI?

This English-language video summarizes the U.S. Copyright Office's 2025 copyrightability report for creators.

As you watch, identify:

Why human authorship remains central
Why prompts alone may not establish authorship of the generated expression
How human modification and arrangement can affect copyrightability
The difference between using AI as a tool and allowing AI to determine the expressive result

Source check: Use the video as an explanation, then consult the U.S. Copyright Office report itself for the authoritative guidance.

2. Case Study: Tweety Bird → An Original Cartoon Canary

Copyright risk can arise when a prompt asks an image model to reproduce a recognizable fictional character. A fictional character may contain protected visual expression, and character names or branding may also raise trademark concerns.

Original Prompt

Prompt: Create an image of Tweety bird singing by the ocean

The prompt directly names an existing fictional character, so the request encourages the system to reproduce recognizable character features.

ChatGPT Image Jul 1, 2025, 08_36_05 PM.png

Direct-reference result: The output closely follows the recognizable character request.

Revised Prompt Used in the Example

Prompt: Create an image of Tweety bird singing by the ocean that will not violate copyright protection laws

This revision communicates the creator's intent, but the prompt still names the protected character. Adding phrases such as “do not violate copyright” is not a legal shield and does not guarantee that the generated image will be sufficiently original.

DALL·E 2026-08-23 22.11.31 - A cute yellow cartoon canary with fluffy feathers, oversized blue eyes, and a playful expression is standing on a rock near the ocean. The canary is s.webp

More original result: The image moves away from the recognizable character design and presents a different cartoon canary.

A Stronger Originality-Focused Prompt

Stronger prompt: Create an original small yellow cartoon canary singing beside the ocean. Give the bird a rounded body, oversized expressive blue eyes, a small orange beak, short orange legs, and outstretched wings. Use a bright, playful storybook aesthetic. Do not reproduce or closely imitate any existing fictional character, character design, signature proportions, facial features, or branded elements.

What Changed?

The protected character name is removed.
The prompt describes general visual attributes rather than asking for a specific character.
The creator defines an original combination of features.
The prompt explicitly asks the system not to reproduce signature character elements.

Copyright lesson: An idea such as “a yellow bird singing by the ocean” is different from asking for the distinctive expression of a particular fictional character. When the protected reference is unnecessary, describe the qualities you need instead of naming the property you want copied.

3. Case Study: Spaceship Earth → Original Futuristic Architecture

Similar concerns can arise when a prompt names a recognizable branded location, building, product, or design. Copyright and trademark questions are separate, and highly recognizable visual properties can raise more than one type of intellectual-property concern.

Original Prompt

Prompt: Create a futuristic home in the year 3025. The image should be based on Disney's Spaceship Earth.

The prompt directly identifies a recognizable Disney landmark as the design reference. The resulting image therefore has a strong incentive to preserve the landmark's distinctive spherical and geometric visual cues.

DALL·E 2025-07-01 20.21.29 - A futuristic home in the year 3025 inspired by Disney's Spaceship Earth. The house has a large geodesic dome structure made of reflective metallic tri.webp

Direct-reference result: The geodesic sphere dominates the design and closely follows the named architectural reference.

Second Prompt Used in the Example

Prompt: Create a futuristic home in the year 3025. The image should be based on Disney's Spaceship Earth, but doesn't violate copyright.

The phrase about copyright communicates intent, but the prompt still tells the model to base the design on the named property. The result can therefore remain visually close to the original reference.

DALL·E 2025-07-01 20.21.36 - A futuristic home in the year 3025 inspired by Disney's Spaceship Earth. The house has a large geodesic dome structure made of reflective metallic tri.webp

Disclaimer-only revision: The wording changes, but the named reference still strongly influences the design.

Translate the Reference Into Original Design Requirements

A stronger strategy is to identify what you actually like about the reference and describe those characteristics without naming the protected or branded property.

Stronger prompt: Create an original futuristic coastal home in the year 3025. Use a large faceted spherical volume integrated into an asymmetrical residence with curved white structural supports, cantilevered glass terraces, indoor gardens, water features, renewable-energy systems, and warm interior lighting. Make the overall silhouette, facade, proportions, entrance, structural layout, and geometric pattern original. Do not reproduce or closely imitate any existing theme-park landmark, branded building, logo, or signature architectural design.

Further Refinement Examples

The following examples move the concept toward a complete residence by adding asymmetry, functional living areas, landscaping, terraces, and a more individualized architectural composition.

ChatGPT Image Aug 23, 2026, 10_14_35 PM (2).png

Refinement A: The sphere becomes one component within a larger asymmetrical coastal residence.

ChatGPT Image Aug 23, 2026, 10_14_35 PM (1).png

Refinement B: Open balconies and asymmetrical supports create a different architectural identity.

Prompt Revision Strategy

Moving from a named reference to an original design brief
Instead of...	Describe...
The name of a protected character, building, or branded property	The functional and visual qualities actually needed for the project
“Make it look like...”	Shape, material, composition, mood, lighting, scale, purpose, and setting
A copyright disclaimer added to the same reference	Originality requirements that change the silhouette, proportions, layout, and signature features

Design lesson: “Inspired by” is not a magic legal category. The safer creative practice is to identify the underlying qualities you want and build a new visual solution rather than asking the model to preserve the recognizable expression of the reference.

4. Disclosure, Provenance & Audience Trust

There is no universal rule requiring every use of generative AI to be disclosed in every context. Disclosure depends on the platform, institution, employer, client, professional field, contract, jurisdiction, and type of content.

Ask Before Publishing

Does my institution or employer have an AI-use policy?
Does the platform require an AI or synthetic-content label?
Could the audience reasonably mistake generated material for a real event, person, quotation, photograph, recording, or experience?
Would disclosure materially help the audience evaluate the content?
Does a client, publication, instructor, grant, or professional standard require documentation of AI assistance?

Example Disclosure Language

Example: Portions of this project were developed with generative AI assistance. Final selection, editing, fact-checking, accessibility review, and creative decisions were completed by the author.

Disclosure should describe what actually happened. Avoid implying that AI was used when it was not, or minimizing its role when the system generated a substantial part of the content.

Platform Disclosure Example: YouTube

YouTube currently requires creators to disclose meaningfully altered or AI-generated content when it appears realistic and could cause viewers to believe that a real person, place, event, or scene was authentic. Non-realistic animation and routine production assistance, such as using AI for outlines or captions, generally do not require the same disclosure.

Content Credentials & Provenance

Technical provenance systems can provide additional information about how media was created. OpenAI-generated images currently include provenance signals such as C2PA Content Credentials and SynthID watermarks.

Provenance can help identify origin, but it does not prove that:

The content is factually accurate
The content has not been edited
The person publishing it owns all necessary rights
The media is being presented in the correct context

Trust principle: Provenance is one signal. It supports transparency, but it does not replace critical evaluation, source verification, or legal review.

5. Ethical Considerations Beyond Copyright

Copyright is only one part of responsible AI content creation. A work can avoid obvious copyright problems and still be misleading, biased, inaccessible, invasive, or inappropriate for the audience.

Accuracy & Misinformation

Verify factual claims, statistics, quotations, names, dates, current events, and sources before publishing. AI-generated realism can increase the risk that inaccurate content will be mistaken for evidence.

Bias & Representation

Review generated content for stereotypes, missing perspectives, unequal representation, assumptions about identity, and visual patterns that repeatedly portray certain groups in limited roles.

Privacy, Likeness & Consent

Do not assume that the technical ability to generate a person's likeness, voice, or personal information means you have permission to use it. Consider privacy, consent, publicity rights, impersonation, and platform rules before creating or publishing realistic synthetic media involving identifiable people.

Authenticity

Do not present invented experiences, testimonials, quotations, photographs, events, or personal stories as authentic. If an AI-generated scenario is fictional, make sure the context does not intentionally mislead the audience.

Accessibility

AI-generated content requires the same accessibility review as manually created content.

Provide meaningful alt text for informative images.
Provide accurate captions for video.
Use readable text and sufficient contrast.
Do not communicate meaning using color alone.
Use descriptive links and logical heading structures.
Review AI-generated descriptions for accuracy before publishing them as accessibility text.

Do Not Treat Detection Tools as Proof

AI detectors and plagiarism checkers can provide signals, but they should not be treated as definitive proof of authorship, originality, infringement, or misconduct. Evaluate the source material and the actual work instead.

Watch & Connect: Responsible AI

This English-language IBM Technology video introduces responsible AI and governance practices that help organizations identify and manage AI risk.

As you watch, identify:

Why ethical AI requires more than technical performance
How governance establishes responsibility
Why transparency and accountability matter
How these principles apply to content-creation workflows

6. A Responsible AI Content-Creation Workflow

Responsible AI use is easier when legal, ethical, accessibility, and quality checks are built into the workflow rather than added after the content is finished.

Create → Review → Verify → Approve

1. Define the Goal
Identify the audience, purpose, platform, format, and level of risk.

2. Check the Inputs
Confirm that source material, images, characters, recordings, and reference content can appropriately be used.

3. Prompt for Originality
Describe the qualities you need rather than asking the model to copy a protected character, property, or distinctive design.

4. Inspect the Output
Look for recognizable protected features, unsupported claims, bias, unwanted likenesses, or misleading realism.

5. Add Human Creative Work
Edit, rewrite, arrange, combine, design, contextualize, and refine the material.

6. Verify & Make Accessible
Check facts, sources, rights, captions, alt text, contrast, readability, and platform requirements.

7. Disclose When Required or Material
Apply applicable institutional, client, legal, professional, and platform requirements.

8. Human Approval
A person makes the final decision to publish, submit, distribute, or use the work.

7. Current Capabilities & Future Directions

Several capabilities once described as the “future of AI” are already part of current content workflows. A more useful question is how these capabilities are becoming more integrated, automated, persistent, and accountable.

What Is Already Here

Current AI capabilities and the direction they suggest for content creators
Current Capability	Direction	Creator Implication
Multimodal creation	Text, image, voice, audio, and visual workflows increasingly operate together.	Creators need cross-media review skills rather than treating each medium separately.
Web search	AI can combine generation with current information and linked sources.	Source literacy and verification become more important, not less.
Projects & persistent context	Long-running work can retain files, instructions, conversations, and project context.	Brand rules, accessibility expectations, and source material can become part of an ongoing workflow.
Plugins & connected tools	AI can move between writing, design, data, research, and external systems.	Permissions, privacy, review points, and governance become part of content strategy.
Provenance signals	Generated media increasingly carries technical signals about origin and creation history.	Creators should understand both what provenance can verify and what it cannot.

Likely Direction of Content Workflows

More connected creation: Fewer handoffs between drafting, research, design, editing, and publishing tools.
More persistent context: Systems will increasingly work from project files, approved terminology, brand standards, and previous decisions.
More agentic workflows: AI systems will complete longer sequences of tasks, increasing the importance of permissions and approval checkpoints.
More provenance and labeling: Platforms and creation tools are continuing to develop ways to communicate how synthetic media was produced.
More governance: Organizations will increasingly need policies defining acceptable AI use, data access, disclosure, review, and accountability.
More emphasis on human differentiation: Original reporting, expertise, first-hand experience, creative judgment, trust, and audience relationships become more valuable as generic content becomes easier to produce.

Future-readiness principle: Do not build a workflow around the assumption that a particular model, interface, or feature will remain unchanged. Build around durable practices: clear goals, strong source material, original human contribution, verification, accessibility, transparency, and accountable review.

8. Final Review Before Publishing AI-Assisted Content

Before publishing AI-assisted content, review the finished product rather than relying on the prompt or the tool's safety language as evidence that the work is ready to use.

Source Rights
Did I have an appropriate basis for using the source and reference material?

Similarity
Does the output reproduce recognizable protected expression, branding, or another person's likeness?

Human Contribution
What creative choices did a person contribute to the final work?

Accuracy
Have factual claims, quotations, dates, citations, and representations been verified?

Privacy & Consent
Does the work contain personal, confidential, biometric, likeness, or voice information that requires permission or additional review?

Bias & Representation
Does the content stereotype, exclude, or misrepresent people or communities?

Accessibility
Are images, links, headings, captions, contrast, language, and multimedia accessible?

Disclosure
Do applicable laws, policies, platforms, contracts, professional standards, or audience needs call for disclosure?

Human Accountability
Is a person prepared to take responsibility for publishing and using the final work?

Try It During Class: Rewrite a Risky Visual Prompt

Choose a visual idea that begins with a named character, branded property, recognizable building, product, or other protected creative reference.

Step 1: Identify the Reference

What specific character, property, design, or branded element is the prompt asking the system to reproduce?

Step 2: Identify What You Actually Need

List the general qualities you want to preserve, such as:

Subject
Shape
Materials
Setting
Mood
Composition
Lighting
Function
Audience

Step 3: Remove the Protected Name

Rewrite the prompt without naming the character, brand, building, artist, franchise, or protected property.

Step 4: Add Originality Requirements

Prompt addition: Make the design original. Do not reproduce or closely imitate the signature appearance, proportions, layout, character design, logo, costume, or other distinctive elements of any existing protected property.

Step 5: Generate & Compare

Compare the result with the original reference and ask:

What recognizable features disappeared?
What features remain too similar?
Does the result have its own identity?
What additional changes would make it more original?

Step 6: Complete the Human Review

Evaluate the final asset for rights, accuracy, accessibility, audience fit, disclosure requirements, and appropriate human creative contribution before using it.

Save This for Later

Copyrightability and infringement are different questions.
Current U.S. copyright guidance continues to require human authorship.
Using AI as a tool does not automatically prevent copyright protection for qualifying human-authored expression.
Prompts alone generally should not be treated as sufficient human authorship of AI-generated expressive elements.
Human-created modifications, selection, arrangement, and other creative expression may be protectable when they meet copyright requirements.
A prompt that says “do not violate copyright” does not guarantee that the output is legally safe.
When a protected reference is unnecessary, describe the visual or functional characteristics you need instead of naming the property.
Review generated images for recognizable characters, branded properties, signature designs, logos, and identifiable likenesses.
Copyright, trademark, privacy, publicity, licensing, and platform-policy questions can overlap.
Disclosure requirements depend on the context; follow applicable institutional, client, platform, professional, contractual, and legal requirements.
Realistic synthetic media may require platform disclosure even when other AI-assisted production tasks do not.
C2PA Content Credentials and watermarking systems can provide provenance signals but do not prove factual accuracy, legal ownership, or correct context.
Responsible AI use also requires attention to bias, misinformation, privacy, consent, authenticity, and accessibility.
Do not use AI detectors or plagiarism checkers as definitive proof of authorship or infringement.
Multimodal creation, web search, persistent projects, connected tools, and provenance systems are current capabilities rather than distant future concepts.
Future content workflows are likely to become more connected and automated, making governance and human approval more important.
Human creators remain responsible for final editorial, ethical, accessibility, and publication decisions.

Reference Resources

U.S. Copyright Office: Copyright and Artificial Intelligence

U.S. Copyright Office: Copyright and Artificial Intelligence, Part 2 — Copyrightability

U.S. Copyright Office: Artificial Intelligence Study

YouTube: Disclosing Use of Generative AI Content

OpenAI: C2PA and SynthID in OpenAI-Generated Images

OpenAI: Verify OpenAI-Generated Content

OpenAI: ChatGPT Search

OpenAI: Projects in ChatGPT

OpenAI: Plugins in ChatGPT and Codex

Matt Wolfe: Here's What You Can and Can't Copyright With AI

IBM Technology: What Is Responsible AI? A Guide to AI Governance