AF7620 Chap.9 Copyright and Ethics of AI in Screen Work
Copyright and Ethics of AI in Screen Work
Human authorship is the bedrock
The production update lecture opens by correcting a common belief: the idea that US law now protects AI art is wrong. It reports that in March 2026 the US Supreme Court refused to take up a leading appeal, so the lower courts' rulings stand: a work produced wholly by artificial intelligence gets no protection. The case, Thaler v.
Perlmutter, concerned a computer scientist who tried to register an image made by his own AI system and named the AI as its sole author.
The US Copyright Office and the courts have held that copyright protects works created by humans, and the courts have called human authorship a bedrock requirement.
Where protection is still possible
A work that an AI produced from a prompt, with no meaningful creative contribution from a person, is refused registration, and merely providing a text prompt is not enough to establish ownership.
Hybrid works that use AI as a tool may still be protected if there is substantial human involvement in editing, arranging or modifying the output, and the Copyright Office examines such works case by case for enough human expression. The lecture notes similar rejections of AI-authored work in the United Kingdom and the European Union, and adds that an applicant for copyright must say whether and how AI was used.
The course's own rule mirrors that duty: when a submission includes AI-generated images, audio or text, you list the prompts and applications used.
Six ethical risks
The opening lecture sets out six issues for users and developers. Ownership of AI-generated work is unclear and involves the fair use of training data. Output can unintentionally copy existing works, and attribution is hard.
AI can produce convincing false content at scale, causing social harm. Bias in training data can produce discriminatory output. Training on personal data raises privacy risks. And legal liability is unsettled: who answers for harm, the developer, the user or the AI?
Two cases from the lecture make the risks concrete: a stock library that sold AI-generated images presented as hand-drawn, which provoked a backlash when a drawing-tablet maker used them, and online investment scams built on deepfake videos.
The lecture also warns that training a fine-tuned adapter on copyrighted material is, in most cases, for personal use only and not for commercial work.
How many jobs are exposed
Survey figures shown in the opening lecture estimate that about 68.7 per cent of firms in US film, television and animation were early adopters of generative AI, and that about 21.4 per cent of jobs in that industry, roughly 118,500, were likely to have enough tasks affected to be consolidated, replaced or eliminated by 2026. For music and sound recording the estimate was 8.4 per cent, about 1,800 jobs, and for gaming 13.4 per cent, about 52,400 jobs.
Roughly one in three film leaders expected displacement for sound editors and 3D modellers. These are forecasts from one survey, and the lecture's own framing is that AI changes tasks unevenly rather than removing whole professions.
What this chapter covers
- 01
Why pure AI art is not protected in the United States
- 02
The Thaler case and human authorship
- 03
Hybrid works and substantial human input
- 04
Disclosing AI use to the Copyright Office and in this course
- 05
Six ethical risks of generated content
- 06
Stock imagery and deepfake cases
- 07
Fine-tuned adapters and copyrighted training data
- 08
Survey estimates of job disruption in film, music and games
Worked example · free
Strengthen the authorship claim of a generated frame
- 2Answer the ownership question accurately.
- 4Propose three changes that add human creative input.
- 2State what must be disclosed.
Key terms
- Hybrid Work
- A work made with AI as a tool but with substantial human involvement in editing, arranging or modifying the output, which may qualify for copyright protection.
- Thaler v. Perlmutter
- The United States case about an image its creator attributed solely to his AI system, in which the courts ruled that a work produced wholly by AI falls outside copyright.
- Deepfake
- A realistic image, video or voice generated or altered by AI to show someone saying or doing something they did not, used in scams and misinformation.
- Training Data
- The collection of examples used to teach an AI model patterns. Its sources raise questions of copyright, fair use, bias and privacy.
- Early Adopter Firm
- In the survey cited in the course, a business that had already implemented or was implementing generative AI programmes in its operations.
- Job Displacement
- The consolidation, replacement or elimination of jobs when enough of their tasks are taken over by a technology.
Copyright and Ethics of AI in Screen Work FAQ
Can I copyright a film made entirely with AI?
Under the United States position described in the lecture, no: a work generated solely by AI from prompts lacks the human authorship copyright requires, and the courts upheld that in the Thaler case. A film in which people wrote the script, directed the choices, arranged and edited the generated material and added their own work is a hybrid, and its human contributions can be protected case by case.
Do I need to tell the teaching team which AI tools I used?
Yes. The teaching plan asks every submission that contains AI-generated images, audio or text to list the prompts and applications used. Keeping a log as you work is the easiest way to comply, and the same record supports any later claim that your contribution was substantial.
Which film jobs are most exposed to generative AI?
In the survey figures shown in the course, film, television and animation leaders most often expected displacement for sound editors and 3D modellers, and about a quarter flagged sound designers, compositors and graphic designers. These are forecasts about tasks within roles, and the course's wider message is that creative design work is assisted by AI more than it is replaced.
Assessment move
Keep a single log for every AI tool you use in the course: date, tool, prompt, what you kept and what you changed by hand. It satisfies the disclosure rule, gives you the evidence of human input a copyright claim needs, and becomes a ready-made appendix for the pipeline strategy in your pitch.
Working through Copyright and Ethics of AI in Screen Work in AF7620? Sia is AskSia’s AI Arts and Humanities tutor — ask any AF7620 Copyright and Ethics of AI in Screen Work question and get a clear, step-by-step explanation grounded in how AF7620 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
