L1
·
Quiz
·
Lab
L2
·
Quiz
·
Lab
L3
·
Quiz
·
Lab
L4
·
Quiz
·
Lab
Module Test
Module 7 · Lesson 1

Copyright, Ownership & AI-Generated Output

Who owns what your AI makes — and what happens when no one does?
If a court rules the AI made it, can you sell it?

The U.S. Copyright Office issued a landmark ruling on the graphic novel Zarya of the Dawn, authored by Kristina Kashtanova. It allowed copyright on the text and arrangement — but revoked protection on every image generated by Midjourney. The ruling stated that AI-generated art "lacks the human authorship necessary to support a copyright claim." Kashtanova had commercially distributed the book before the ruling. She was legally exposed.

The Ownership Vacuum

Copyright law in virtually every major jurisdiction — the U.S., EU, UK, Australia — requires a human author. When an AI model produces an image, a piece of music, or a block of code without meaningful human creative expression in the specific output, that output enters the public domain the moment it is created. Anyone can copy, resell, or remix it. Your competitive advantage evaporates.

This is not theoretical risk. The U.S. Copyright Office has now issued multiple decisions on AI content. In March 2023 it released formal guidance: AI-generated material where a human did not "select and arrange" the elements with sufficient creativity is unprotectable. In 2024 it denied registration for an AI-generated image submitted by artist Jason Allen — whose piece "Théâtre D'Opéra Spatial" had won a Colorado State Fair art competition — despite his extensive prompting effort.

Why This Hits AI Side Hustlers Hard

If you sell AI-generated stock images, illustrations, book covers, or music through platforms like Adobe Stock, Shutterstock, or DistroKid, you may be reselling material you cannot legally protect. Several stock platforms already ban purely AI-generated uploads. Others require disclosure. Selling without disclosure may violate platform terms and expose you to account termination and clawback of earnings.

Where Human Authorship Survives

Copyright does attach when a human author uses AI as a tool while exercising sufficient creative control over the final output. The Copyright Office has acknowledged this in rulings on comic book layouts and hybrid works. The practical standard emerging from decisions in 2023–2024: did a human make expressive choices about selection, arrangement, and creative expression — not just type a prompt?

Strategies that strengthen the human authorship argument include: heavily editing AI output in Photoshop or Illustrator, writing the prompt as part of a larger curated creative process, combining AI output with original photography or hand illustration, and documenting your creative process for registration purposes.

Practical Risk Map
High Risk
Prompt-only AI images sold as stock
Minimal human creative expression. No copyright. Anyone can legally copy your exact upload after purchase.
High Risk
AI-written articles sold as ghostwritten originals
Client may own nothing. If they publish and another AI user generates the same text, neither has a claim.
Lower Risk
AI-assisted book with heavy human editing
Human author makes expressive choices throughout. Copyright attaches to the human-authored portions.
Lower Risk
AI code + human architecture & documentation
Software copyright protects the overall structure and original elements contributed by the human developer.
Key Terms
Work for HireA work created by an employee or under contract where the employer/commissioner owns the copyright. AI cannot be an employer, so "work for hire" doctrine does not apply to AI output itself.
Public Domain by DefaultWhen no human authorship exists, AI output is immediately public domain — no registration required to lose protection, because protection never attached.
Sufficient Creative ExpressionThe USCO standard: a human must contribute original, expressive elements beyond a mere functional prompt for copyright to attach to specific output.
Practical Takeaway

Register your copyright only on works where you have documented human creative contribution. Disclose AI involvement to clients and platforms honestly. Build your business value around your curation, editing, and process — not the raw AI output — because that is what the law actually protects.

Quiz — Copyright & AI Ownership

3 questions · Select the best answer
1. The U.S. Copyright Office's 2023 guidance states that AI-generated content is:
Correct. The USCO requires human authorship — specifically original expressive choices — for copyright to attach. Prompt length is irrelevant to the legal test.
Incorrect. There is no AI copyright category. The USCO requires human creative expression. Without it, AI output enters the public domain immediately.
2. In the Zarya of the Dawn ruling, what happened to the Midjourney-generated images?
Correct. The USCO revoked image copyright, finding the Midjourney outputs lacked the human authorship required by law. Only her text and arrangement remained protected.
Incorrect. Payment for an AI service does not create copyright. The USCO revoked image protection specifically because no human authorship existed in the generated images.
3. Which strategy best strengthens a copyright claim on AI-assisted work?
Correct. Documented human creative expression — editing choices, arrangement decisions, original additions — is the legal foundation for copyright in AI-assisted work.
Incorrect. Prompt complexity and publication speed do not create copyright. Human creative contribution to the final expressive output is what the law requires.

Lab 1 — Copyright Scenario Advisor

Practice evaluating real copyright risk in AI side hustle scenarios

Your Task

Describe a specific AI side hustle scenario — a product you plan to sell, a service you're offering, or a type of content you're creating with AI — and the advisor will help you assess your copyright exposure and what steps would strengthen your legal position.

Try: "I'm selling AI-generated children's book illustrations on Etsy. The images come directly from Midjourney prompts. What's my copyright situation?" — or describe your own scenario.
Copyright Risk Advisor
AI-Assisted
Welcome. Describe your AI content business or product — what you're creating, how you're creating it, and how you plan to sell or distribute it. I'll help you assess your copyright exposure and identify what steps strengthen or weaken your legal position under current U.S. Copyright Office guidance.
Module 7 · Lesson 2

Disclosure, Deception & Client Trust

When does not mentioning AI become fraud — and what actually happened when it did?
At what point does using AI without telling your client cross a legal line?

A New York law firm, Levidow, Levidow & Oberman, submitted a legal brief in Mata v. Avianca Airlines citing multiple court cases that did not exist. The firm's attorney Steven Schwartz had used ChatGPT to research the brief and submitted AI-fabricated citations without verifying them. When the court discovered the nonexistent cases, it sanctioned the firm $5,000 and required all attorneys involved to complete AI training. The case became the defining example of undisclosed AI use causing concrete professional harm.

The Deception Problem

Most freelance and service contracts contain an implicit or explicit representation that the deliverable reflects the seller's own skill and labor. When clients pay a premium for "expert" copywriting, legal research, financial analysis, or creative work, and receive AI output passed off as original professional work, the gap between representation and reality can constitute fraud or breach of contract under common law in most U.S. states.

The FTC Act Section 5 prohibits "unfair or deceptive acts or practices." In 2023, the FTC issued guidance specifically noting that AI-enabled deception — including falsely representing AI-generated content as human-created — falls within its enforcement authority. While the FTC has focused on large platforms in early AI enforcement actions, the legal framework applies to individual sellers.

The Schwartz Standard

The Mata v. Avianca sanctions established that professionals who use AI without verification and without disclosure — in contexts where accuracy and human judgment are material to the engagement — face professional discipline, client liability, and reputational destruction. The standard is not specific to lawyers; it applies wherever a client reasonably relies on professional expertise.

Platform Disclosure Rules (2023–2024)

Several major platforms have implemented mandatory AI disclosure requirements. Upwork updated its terms in 2023 to require freelancers to disclose AI tool use when clients ask, and prohibits delivering AI output as the freelancer's own original work when the client hasn't consented to AI use. Fiverr added AI disclosure fields to gig listings. Academic Influence and similar content buyers explicitly ban AI content in their contractor agreements.

The EU's AI Act (formally adopted 2024) imposes disclosure obligations on AI-generated content in certain contexts — particularly content that could be mistaken for human-created work in consequential domains. UK and Australian regulatory frameworks are following similar trajectories.

A Practical Disclosure Framework
  • Before starting any client engagement, ask: "Does this client believe they are paying for human-only work?" If yes, disclose upfront or don't use AI.
  • Include a short AI disclosure clause in your contract: "Deliverables may incorporate AI-assisted drafting, which [Client/Contractor] reviews and edits for accuracy and quality."
  • Never submit AI output that you have not verified for factual accuracy — especially in legal, medical, financial, or technical contexts.
  • If a client contract explicitly prohibits AI tools, honor it. Using AI anyway creates breach of contract exposure regardless of output quality.
  • On platforms with disclosure fields, fill them accurately. Platform terms violations can result in account bans and payment clawbacks.
Where Disclosure Protects You

Transparent AI use, disclosed upfront, is not a legal liability — it is a service feature. Many clients actively want AI-assisted work for speed and cost. The risk is not AI use itself but misrepresentation of what the client is receiving. A contract that accurately describes AI-assisted deliverables, with the freelancer's curation and quality control as the core service value, is legally sound and increasingly common.

The Rule of Thumb

If you would be uncomfortable telling your client exactly how you produced their deliverable, that discomfort is the signal you need to disclose — or reconsider how you're working. Transparency is not weakness; in 2024, it is increasingly the professional standard that separates reputable AI-assisted service providers from those who face legal exposure.

Quiz — Disclosure & Deception

3 questions · Select the best answer
1. In Mata v. Avianca, attorney Steven Schwartz was sanctioned primarily because:
Correct. The court found Schwartz submitted fabricated citations he had not verified, in a context where accuracy and professional judgment were material. The lack of verification and disclosure — not AI use itself — drove the sanctions.
Incorrect. Courts have not categorically banned AI research tools. The sanctions arose from submitting unverified, AI-hallucinated citations as real cases — a failure of professional duty and accuracy.
2. Under Upwork's 2023 updated terms, a freelancer who delivers AI-generated work as their own original output without client consent:
Correct. Upwork's updated terms treat undisclosed AI-as-original-work as a terms violation. Positive client feedback does not retroactively legitimize misrepresentation of the work's origin.
Incorrect. There is no earnings percentage penalty or dollar threshold in Upwork's terms. Misrepresenting AI work as original human work violates terms regardless of contract size or client satisfaction.
3. Which approach best protects a freelancer who uses AI tools in their work?
Correct. Upfront contractual disclosure and rigorous output verification are the two key protections. They prevent misrepresentation claims and eliminate the professional liability that arises from unverified AI output in consequential work.
Incorrect. Silence about AI use — regardless of contract value or output quality — creates ongoing legal and reputational risk. Disclosure and verification are the appropriate protections.

Lab 2 — Disclosure Language Drafter

Write real contract language and client communications for AI-assisted services

Your Task

Tell the advisor what type of freelance service you offer and what AI tools you use. It will help you draft clear, professional contract disclosure language and client-facing communications that protect you legally while positioning AI use as a feature, not a liability.

Try: "I write SEO blog posts for local businesses using Claude. What should my contract say about AI?" — or describe your own service type.
Disclosure Language Drafter
AI-Assisted
Tell me about your service: what do you deliver, what AI tools do you use, and what platform or context do you work in (direct clients, Upwork, Fiverr, etc.)? I'll help you draft disclosure language for your contracts and client communications.
Module 7 · Lesson 3

Training Data, Defamation & Output Liability

When AI says something false about a real person — who gets sued?
If your AI tool fabricates a defamatory statement and you publish it, are you liable?

Radio host Mark Walters sued ChatGPT maker OpenAI in Georgia state court after journalist Fred Riehl used ChatGPT to summarize a legal complaint, and ChatGPT fabricated accusations that Walters had embezzled funds from a gun rights organization. The accusations were entirely false — no such complaint existed. Walters v. OpenAI became the first U.S. defamation lawsuit against an AI company over hallucinated content. The case raised the question courts had not yet answered: does a publisher who relies on AI output bear the same liability as a traditional publisher?

The Hallucination Liability Chain

AI hallucinations — confident fabrications presented as fact — create a liability chain. The AI developer may have Section 230 protection in the U.S. as a platform intermediary, though this remains actively litigated. The person who publishes the AI output, however, is a traditional publisher under existing defamation law and generally does not benefit from the same protections.

If you use an AI tool to write content about real, identifiable people — reviews, news summaries, profiles, research reports — and that content contains false statements of fact, you face potential defamation liability under the law of the jurisdiction where the subject resides. Publication on a website accessible in that jurisdiction is typically sufficient to establish personal jurisdiction.

The Training Data Dimension

The New York Times filed suit against OpenAI and Microsoft in December 2023, alleging that training on Times articles without license constitutes copyright infringement. The lawsuit included examples where GPT-4 reproduced Times articles nearly verbatim. If you resell AI content that substantially reproduces a third party's copyrighted training data, you may be in the downstream liability chain — particularly if you knew or should have known the output matched protected source material.

Defamation Risk in AI Side Hustles

The highest-risk AI side hustle activities for defamation exposure include: AI-generated reviews or ratings of specific businesses, AI-generated profiles or bios of real individuals, AI-written news summaries or investigative content, and AI-assisted competitive intelligence reports mentioning named companies or executives.

Lower-risk activities: generic educational content, fictional characters, product descriptions for your own products, and abstract how-to content with no named real-world subjects.

Privacy Law: The GDPR & CCPA Dimension

If you use AI to process, analyze, or generate content about individuals' personal data — particularly EU residents (GDPR) or California residents (CCPA) — you may have data controller obligations. In August 2023, Italy's data protection authority temporarily banned ChatGPT over GDPR compliance concerns. Using AI tools to scrape, compile, or analyze personal data about individuals as a commercial service triggers these obligations regardless of where you are located.

Reducing Output Liability
  • Verify every factual claim about real, named individuals before publishing — AI output about specific people is the highest hallucination-liability intersection.
  • Add a content review step to every workflow involving real-world facts: treat AI output as a first draft, never a final publication.
  • Avoid AI content about private individuals by name; public figures have a higher legal threshold for defamation but the reputational risk still applies.
  • Keep records of your editorial review process — documentation that a human reviewed and verified content before publication is a meaningful defense in negligence-based claims.
  • If your service involves personal data of EU or California residents, consult a privacy attorney about GDPR/CCPA compliance obligations for your specific use case.
The Bottom Line on Output Liability

You are the publisher. AI is your tool. Publishers are legally responsible for what they publish under defamation, copyright, and privacy law — and "the AI wrote it" has not been accepted as a defense in any court ruling to date. Build a verification habit into every workflow that touches real facts and real people.

Quiz — Defamation & Output Liability

3 questions · Select the best answer
1. Walters v. OpenAI is significant because it:
Correct. Walters v. OpenAI was the first U.S. defamation lawsuit specifically targeting an AI company for fabricated statements about a real individual — setting a legal precedent for AI hallucination liability claims.
Incorrect. The case did not result in a ruling that AI companies have blanket immunity or that Section 230 fully covers all AI output. It was notable as the first such defamation suit, raising unresolved liability questions.
2. Under current defamation law, if you publish AI-generated content that falsely accuses a named person of a crime:
Correct. As the publisher, you bear legal responsibility for what you publish. No court has accepted "the AI generated it" as a complete defense to defamation. The AI is your tool; you are the publisher.
Incorrect. Publishers — not tools — bear defamation liability. Charging money is irrelevant to defamation liability. The act of publishing the false statement about a real person creates exposure regardless of origin.
3. Which AI side hustle activity carries the HIGHEST defamation and output liability risk?
Correct. AI-generated content about real, named individuals or businesses is the highest-risk category — combining hallucination risk with direct defamation exposure when false statements of fact are published about identifiable people.
Incorrect. Fictional characters, generic product descriptions, and how-to summaries carry far lower defamation risk because they typically don't make false factual claims about real identifiable people.

Lab 3 — Content Liability Checker

Evaluate specific AI content for defamation and publication risk before you publish

Your Task

Share a sample of AI-generated content you're considering publishing, or describe the type of content your AI side hustle produces. The advisor will flag specific liability risks — defamation, hallucination exposure, privacy issues — and suggest how to mitigate them before publication.

Try: "I'm creating AI-written business profiles for a local directory website. Here's an example: [paste sample]" — or describe your content type and I'll walk through the risks with you.
Content Liability Checker
AI-Assisted
Share the AI-generated content you're evaluating, or describe what you're creating and who it's about. I'll help you identify specific defamation, hallucination, and privacy risks — and tell you exactly what to check or change before you publish.
Module 7 · Lesson 4

Building an Ethically Defensible AI Business

The gray areas are real — and the practitioners who navigate them well will build lasting businesses.
What does a responsible, sustainable AI side hustle actually look like in practice?

When the writers' strikes concluded in September 2023, the WGA secured the first major industry contract provisions governing AI use — including the right to know if AI was used in developing projects they're hired to rewrite, and protections against having their work used to train AI models. The SAG-AFTRA AI agreement in 2024 established consent and compensation requirements for synthetic voice and likeness use. These were not theoretical demands. They reflected the market realities that AI freelancers operate in: clients and counterparts now have explicit contractual expectations about AI disclosure.

The Regulatory Trajectory

The EU AI Act, formally enacted in 2024, imposes mandatory transparency requirements, prohibitions on certain AI uses, and compliance obligations that will affect AI-based service providers selling into European markets. The act classifies AI applications by risk level — certain high-risk uses (employment screening, credit scoring, biometric identification) face strict conformity requirements; general-purpose AI systems face transparency and registration obligations.

U.S. federal regulation remains fragmented, but 2023–2024 saw executive orders on AI safety, FTC guidance on AI deception, and state-level legislation in Colorado, Connecticut, and California specifically addressing automated decision-making transparency. The direction is toward more disclosure requirements, not fewer.

The Ethical Gray Areas That Actually Arise

Beyond legal compliance, practitioners encounter genuine ethical gray areas where the law is silent but professional reputation is at stake. The most common in AI side hustles:

Gray Area
Ghostwriting with AI for individuals
Ghostwriting has always existed. AI makes it faster and cheaper. The question is whether the client understands what they're receiving and whether the platform or publisher prohibits AI-assisted ghostwriting.
Gray Area
AI voice cloning for content creators
ElevenLabs and similar platforms require consent for cloning real voices. Using a cloned voice commercially without consent may violate right-of-publicity laws in many states and the SAG-AFTRA 2024 agreement.
Defensible
Disclosed AI-assisted research & writing
Clients who consent to AI-assisted work, with human review and accuracy guarantees, are the professional standard emerging across industries. This is where sustainable business is built.
Defensible
AI automation with transparent pricing
Charging for speed, volume, and curation rather than for "human hours" that no longer apply is honest and increasingly expected. Clients who value this are the right clients.
The Five-Point Ethical Framework
  • Accuracy first: Every AI output touching real facts requires human verification before delivery. No exceptions for "small" engagements.
  • Consent always: Clients, subjects of content, and platform terms must all be aligned with how you're using AI in the work.
  • Transparency by default: Default to disclosure. Opt out only when a client explicitly instructs otherwise and the law permits it.
  • Value on curation: Price and position your service on what you uniquely contribute — judgment, taste, expertise, editing — not on the AI output itself.
  • Documentation as protection: Keep records of your process — prompts used, edits made, verification steps taken. Documentation is your defense in any legal or professional dispute.
What Clients Are Actually Buying

The freelancers who thrive in AI-saturated markets understand a fundamental shift: clients are no longer paying primarily for the labor of producing content. They are paying for judgment — the human expertise to know what's good, what's accurate, what's appropriate, and what will actually achieve their goal. AI lowers the cost of production; it does not replace the value of expert judgment about what to produce.

The AI practitioners who build lasting businesses are those who document their expert process, deliver verified output, disclose AI involvement honestly, and price on the value of outcomes — not the hours of labor that AI has compressed. That is both the ethical and the commercially sound approach to operating in this space.

Module 7 Core Principle

Legal gray areas exist because the law has not yet caught up with AI capabilities. In that space, your reputation and your clients' trust are your most valuable assets — more valuable than any individual project. The ethical framework here is not idealism; it is long-term business strategy. The practitioners who are still operating in five years will be the ones who got this right from the start.

Quiz — Ethical & Defensible AI Practice

3 questions · Select the best answer
1. What key provision did the WGA secure in their 2023 strike settlement regarding AI?
Correct. The WGA agreement secured disclosure rights and training data protections — not a ban on AI. This reflects the emerging professional standard: consent, transparency, and protection of human creative work from unauthorized AI training use.
Incorrect. The WGA did not ban AI tools. They secured transparency and consent provisions — knowing when AI is used and protecting their work from being used to train models without consent.
2. Using AI voice cloning commercially to produce content with a real person's cloned voice without their consent most likely violates:
Correct. Right-of-publicity laws protect individuals' control over commercial use of their identity including voice. The SAG-AFTRA 2024 agreement codified consent requirements for AI voice and likeness use in relevant industries.
Incorrect. Right-of-publicity laws — which exist in most U.S. states — protect against unauthorized commercial use of a person's voice and likeness. This is not limited to television broadcasting or platform terms alone.
3. In an AI-saturated freelance market, what do clients primarily pay for that AI cannot replace?
Correct. AI compresses production labor; it does not replace expert judgment. Clients who understand AI know they are paying for your expertise in knowing what to prompt, what to verify, what to edit, and what to deliver — not the hours AI saved.
Incorrect. AI has largely commoditized tool access, and clients can access AI tools themselves. The durable value is human judgment — accuracy verification, strategic direction, and expert curation — that AI cannot reliably provide.

Lab 4 — Ethical Business Stress-Tester

Map your AI side hustle against the five-point ethical framework

Your Task

Describe your current or planned AI side hustle in as much detail as you're comfortable sharing. The advisor will stress-test it against the five-point ethical framework from Lesson 4 — accuracy, consent, transparency, value positioning, and documentation — and identify the specific gray areas you need to navigate for a legally and ethically defensible business.

Try: "I'm building an AI-powered content agency that writes LinkedIn posts and newsletters for SaaS founders. Here's how I work: [describe your process]" — or describe your own business model.
Ethical Business Stress-Tester
AI-Assisted
Describe your AI side hustle: what you do, how you use AI in your process, what you deliver, and how you position it to clients. I'll evaluate it against the five-point ethical framework — accuracy, consent, transparency, value positioning, and documentation — and flag any specific gray areas you should address.

Module 7 — Test

15 questions · 80% required to pass · Legal and Ethical Gray Areas
1. The U.S. Copyright Office's position on purely AI-generated output is that it:
Correct. Without human authorship, AI output is immediately public domain. No copyright attaches at creation.
Incorrect. The USCO requires human creative expression. Without it, output is public domain from the moment of creation.
2. In the Zarya of the Dawn case, the Copyright Office:
Correct. The split ruling is the key precedent: human creative expression is protected; AI image output without sufficient human authorship is not.
Incorrect. The ruling was split — text and arrangement protected, AI images not. Using AI tools does not automatically strip all copyright from a hybrid work.
3. Which activity creates the strongest copyright claim for an AI-assisted creator?
Correct. Documented human creative contribution — editing, selection, original additions — is what the USCO standard requires for copyright to attach to AI-assisted work.
Incorrect. Prompt complexity, unedited AI output, and registration without human creative expression do not create copyright protection under current USCO guidance.
4. What was the primary error that led to the Mata v. Avianca sanctions against attorney Schwartz?
Correct. Schwartz submitted hallucinated citations as real cases without verification. The failure was not AI use itself but the absence of the verification step that professional duty requires.
Incorrect. No ABA prohibition on ChatGPT exists. The sanctions arose from submitting unverified fabricated citations as real cases to a court — a fundamental professional failure.
5. Under FTC Act Section 5, falsely presenting AI-generated content as human-created to clients:
Correct. The FTC's 2023 guidance explicitly noted that AI-enabled misrepresentation falls within its Section 5 authority — including misrepresenting AI output as human-created work.
Incorrect. Section 5 applies to individuals as well as corporations, covers all content types, and is not limited to resale. FTC guidance specifically addressed AI deception in 2023.
6. Upwork's 2023 updated terms require freelancers to:
Correct. Upwork's terms center on disclosure and consent — not tool restrictions, surcharges, or platform review of deliverables.
Incorrect. Upwork does not impose surcharges, platform review, or approved tool lists. The requirement is consent-based disclosure when clients inquire about AI use.
7. The most effective protective clause to add to a freelance contract when using AI tools is:
Correct. An accurate AI disclosure clause with a human review representation protects against misrepresentation claims while being honest about your process.
Incorrect. False guarantees, copyright waivers, and AI company indemnification (which AI companies do not provide) are all ineffective or counterproductive protective strategies.
8. Walters v. OpenAI raised which previously unresolved legal question?
Correct. The case raised the unresolved question of AI company liability for hallucinated defamatory content and whether Section 230 platform protections apply to AI output.
Incorrect. The case concerned defamation liability for AI hallucinations and Section 230's applicability — not patent law, journalist exemptions, or First Amendment protection for AI content.
9. If you publish AI-generated content that falsely states a named individual committed fraud, under defamation law you are:
Correct. You are the publisher. Publishers bear defamation liability regardless of their production tool. AI authorship has not been accepted as a complete defense in any ruling to date.
Incorrect. Neither AI authorship nor generic disclaimers constitute a complete defense to defamation. Publisher liability applies to you as the person who chose to publish the false statement.
10. The New York Times lawsuit against OpenAI and Microsoft (December 2023) alleged:
Correct. The Times lawsuit centered on training data copyright — arguing that training on its articles constituted infringement, with GPT-4's near-verbatim reproduction as evidence.
Incorrect. The case was about training data copyright infringement, not defamation, brand impersonation, or privacy policy violations.
11. Italy's data protection authority temporarily banned ChatGPT in 2023 primarily over:
Correct. Italy's Garante authority temporarily banned ChatGPT over GDPR compliance — specifically how personal data of users and subjects was collected, processed, and stored.
Incorrect. The ban was GDPR-based, not related to defamation, copyright, or national security concerns specifically.
12. The EU AI Act (enacted 2024) classifies AI applications by:
Correct. The EU AI Act's risk-based framework is its defining structure — high-risk uses face strict requirements; general-purpose systems face transparency and registration obligations.
Incorrect. The Act applies based on use risk, not developer nationality, revenue threshold, or content type. Text generation is not exempt.
13. The WGA's 2023 contract provisions on AI established:
Correct. The WGA secured disclosure rights and training data protections — reflecting the real-world consent and transparency standards now embedded in professional creative industry contracts.
Incorrect. The WGA did not ban AI, set wage increases for AI skills, or require AI script credits. The provisions centered on disclosure and protection from unauthorized training data use.
14. Using a real person's cloned AI voice commercially without their consent most directly risks violating:
Correct. Right-of-publicity laws protect commercial use of an individual's voice and likeness. The SAG-AFTRA 2024 AI agreement codified consent requirements for the entertainment industry specifically.
Incorrect. State right-of-publicity laws remain active (they were not preempted), FCC rules are irrelevant to voice cloning per se, and platform terms are not the only legal constraint.
15. In an AI-saturated freelance market, the sustainable business positioning is:
Correct. AI has commoditized content production. The durable competitive advantage is expert judgment — knowing what to produce, verifying it's accurate, and curating it effectively — which AI alone cannot reliably provide.
Incorrect. Hiding AI use creates legal and reputational risk; pure price competition is a race to the bottom; and tool operation without domain expertise is itself commoditized. Human judgment and expertise are the sustainable differentiators.