AI, short for artificial intelligence, doesn't have one clear-cut definition. According to the Canadian government, AI refers to systems that use data to automate complicated tasks and occasionally make decisions or predictions about people.

The main goal of AI developers is to create machines capable of learning, understanding intentions, performing tasks, and making decisions similar to humans. It operates through algorithms by dealing with probabilities and relying on extensive training from data to provide the best possible approximation.

AI includes various technologies such as machine learning, deep learning, neural networks, large language models, expert systems, and robotics.

Generative AI is a part of artificial intelligence that can make new things like text, images, audio, or code. It can also alter existing things like recorded language, accents, facial expressions or movements. It learns from the patterns and structure of the input data to create outputs that look like they were made by humans.

AI is being used more and more in various fields like voiceover, film, TV, digital media, music, dance, and gaming. Although AI isn't widely used yet to create audio and visual content, it's developing quickly, and it won't be long before it becomes common.

ACTRA Toronto is specifically worried about the "digital double," where machine learning systems create a synthetic performance by recording or reproducing a performer's voice, persona, or likeness. This technology has many applications, including:

• Translating or generating text-to-voice or image-to-voice;
• Creating interactive digital humans or avatars that can interact with users; and
• Manipulating existing identities in audiovisual content, commonly known as deepfakes.

As of now, AI falls short in reproducing the creativity, emotion and authenticity Performers bring to a performance. Nevertheless, AI can serve as a helpful tool in the creative process, with humans retaining the ultimate decision-making authority. In specific areas of entertainment, however, AI has begun to imitate certain aspects of Performers' roles, a trend expected to grow as the technology advances.

If AI is applied ethically and responsibly in collaboration with workers, their union and the government, it has the potential to positively impact Performers. For example, AI could allow Performers to appear in multiple productions around the world across a single period, boosting income levels. AI could increase accessibility to the labour market for Deaf and disabled Performers, while enhancing the safety of working as a Stunt Performer.

• Pay - The one-off payments offered to Performers who engage in AI work often do not reflect the fact that their image, voice or likeness may be used in perpetuity and on thousands of different projects.
• Rights - Many generative AI platforms are using original creative work without a license to create new material. This is infringing on ACTRA members’ intellectual property (personal voices, sound effects, actions, behaviour, images, likenesses and personalities) and legal (consent) rights as well as their interfering with their ability to earn a viable living.
• Transparency - Performers often do not know where or how their professional contribution is being exploited due to unclear contract provisions, unspecified usage and a lack of auditing.
• Awareness - When reviewing their contract or rider, Performers should be aware of accepting language, such as simulation, synthetization, digital double and machine learning, without first understanding the potential consequences of how their consent will impact the use of their image/likeness/voice and future compensation streams.
• NDAs - Performers are being asked to sign excessive Non-Disclosure Agreements without any knowledge of what the job entails, which is far too common in the industry.
• Reputation - Unlike most professions, the majority of Performers exist in a completely public-facing line of work with their image, likeness and voice on display for the world to see. When their image, likeness or voice is used without their consent, especially for nefarious purposes, it can cause irretrievable reputational harm. The unregulated use of AI increases the chances of a performer’s work being defamed, edited or otherwise modified by AI without their permission.
• Artistic Expression – A performer’s artistic output is an extremely personal yet collaborative process. The increasing use of AI-generated content is making it more difficult for audiences to distinguish between AI-generated content and a performer's own work. The use of AI-generated content is not only being used as a substitute for human creativity, but it is impeding on a performer’s right to artistic expression. Additionally, AI-generated content is often used without the performer’s consent or financial compensation.
• Harmful content - It is estimated that 96% of deepfakes depict women and are pornographic, and 99% of deepfake subjects are from the entertainment industry. (source: Sensity AI 2019 report)
• Equality - Rapid technological developments have the potential to worsen existing inequalities in access to work, income inequalities and wage gaps for individuals from marginalized communities across the industry. These systems scrape data from pre-existing biased sources, which can lead to deepening racial and economic inequalities. (source: ACLU report)
• Collective bargaining - This technology is challenging for Performers’ unions globally as this innovation would not be covered in historic collective bargaining agreements.
• Canadian law - There are no current Canadian laws explicitly written to regulate AI. This includes the Copyright Act, which provides the legal framework for Performers’ intellectual property rights. This is leading to some Performers being exploited. ACTRA has submitted recommendations to the Standing Committee on Industry and Technology’s Artificial Intelligence and Data Act consultation to safeguard the fundamental needs of Canadian performers: respect, fair wages, and protection against abuse.

READ: The Canadian Intellectual Property Office: A Guide to Copyright

Copyright applies to every original literary, dramatic, musical and artistic work created by the author/creator using their own skill and judgment. It also applies to other subject-matter consisting of performer's performances, sound recordings and communication signals. For these, the applicable rights may differ somewhat. The following are protected by copyright when done by a performer:

• a performance of an artistic, dramatic or musical work, whether or not the work was previously fixed (recorded) and whether or not the work's term of copyright protection has expired;
• a recitation or reading of a literary work, whether or not the work's term of copyright protection has expired; and,
• an improvisation of a dramatic, musical or literary work, whether or not the improvised work is based on a pre-existing work.
• recorded performances of actors, authors, singers, musicians and dancers on tapes, cassettes, CDs, CD-ROMs, video recordings and films, compilations of performances by performers on records, CDs and in audiovisual formats.

Copyright lasts until the end of 50 years after the end of the calendar year in which the performance occurs.

If the performance is fixed in a sound recording before the copyright expires, the copyright continues for 50 years after the end of the calendar year in which it is first fixed.

If the sound recording is published before the copyright expires, the copyright continues until 50 years after the end of the calendar year in which the first publication occurs or 99 years after the end of the calendar year in which the performance occurs, whichever is earlier.

Who owns the copyright?

Where the subject matter is something other than a work of literature or art, the copyright owner is usually:

• The maker (the person or organization that made the arrangements for the first fixation of the sounds) is the first owner of the copyright in a sound recording.
• The performer is the first owner of the copyright in a performer’s performance.
• The broadcaster is the first owner of the copyright in a communication signal.

While a performer may own their performance, they do not own the rights involved in the broadcast and distribution of that performance.

ACTRA Toronto is asking the federal government to step-in and ensure AI system developers and users are held liable for complying with existing legal frameworks, including intellectual property, data protection and privacy laws. As a longer-term priority, we are campaigning for the Artificial Intelligence and Data Act (AIDA) to ensure wholesale reform of our intellectual property framework, including new Consent, Compensation and Control rights for Performers.

In early 2024, the Australian Government released an interim response following a consultation in 2023. The interim response outlines proposed next steps for AI reform and focuses on the following key areas:

1. preventing harms from occurring through testing, transparency and accountability;
2. clarifying and strengthening laws to safeguard citizens;
3. working internationally to support the safe development and deployment of AI; and
4. maximizing the benefits of AI.

As part of the initiatives proposed for addressing point 1 above, the Government committed to developing an AI Safety Standard and implementing risk-based guardrails for the industry, which were released on September 5, 2024, as part of its measures to regulate AI.

There are 10 voluntary guardrails (which form part of the Voluntary AI Safety Standard) currently in effect that apply to all Australian organizations and all organizations throughout the AI supply chain as well as 10 mandatory guardrails, applicable to AI in high-risk settings, that have been released in draft form and are subject to further consultation.

Brazil has proposed a comprehensive AI Bill, which is currently being debated (as of July 2024) by the Brazilian government.

The Bill’s primary aim is to grant individuals significant rights and place specific obligations on companies that develop or use AI technology (AI supplier or operator). To achieve this, the Bill establishes the creation of a new regulatory body to enforce the law and takes a risk-based approach by categorising AI systems into different categories. It also introduces a protective system of civil liability for providers or operators of AI systems, along with a reporting obligation for significant security incidents.

The Government of Canada tabled the Artificial Intelligence and Data Act (AIDA) in June 2022 as part of Bill C-27, the Digital Charter Implementation Act, 2022. Following the second reading in the House of Commons in April 2023, Bill C-27 is currently being studied by the Standing Committee on Industry and Technology.

The AIDA proposes the following approach:

• Building on existing Canadian consumer protection and human rights law, AIDA would ensure that high-impact AI systems meet the same expectations with respect to safety and human rights to which Canadians are accustomed. Regulations defining which systems would be considered high-impact, as well as specific requirements, would be developed in consultation with a broad range of stakeholders to ensure that they are effective at protecting the interests of the Canadian public, while avoiding imposing an undue burden on the Canadian AI ecosystem.
• The Minister of Innovation, Science, and Industry would be empowered to administer and enforce the Act, to ensure that policy and enforcement move together as the technology evolves. An office headed by a new AI and Data Commissioner would be created as a centre of expertise in support of both regulatory development and administration of the Act. The role would undergo gradual evolution of the functions of the commissioner from solely education and assistance to also include compliance and enforcement, once the Act has come into force and ecosystem adjusted.
• Prohibit reckless and malicious uses of AI that cause serious harm to Canadians and their interests through the creation of new criminal law provisions.

The Government of Canada launched the Canadian Artificial Intelligence Safety Institute (CAISI) on Tuesday, November 12, 2024, to leverage Canada’s world-leading AI research ecosystem and talent base to advance the understanding of risks associated with advanced AI systems and to drive the development of measures to address those risks. CAISI will conduct research under two streams: Applied and investigator-led research; and Government-directed projects.

China does not have a comprehensive AI Act but does have a number of regulations that focus on subsets of AI.

China’s Interim Measures for the Management of Generative Artificial Intelligence Services came into effect on August 15, 2023. The AI Measures are formulated to "promote a healthy development and regulated application of generative artificial intelligence, safeguard national security and social public interests, and protect the lawful rights and interests of citizens, legal persons and other organizations.

The Measures include:

• Lawful use
• Data labeling rules
• Data training
• Content moderation
• Reporting mechanism

The AI Measures apply to companies when they provide Generative AI services to the public within China regardless of where they are incorporated.

The Deep Synthesis Provisions, which came into force in January 10, 2023, are designed to implement the Outline for the Construction of a Rule of Law-governed Society (2020-2025) formulated by the Chinese central government. The goal is to improve the standardized management of new technologies such as algorithmic recommendations and deep fakes, define the "bottom line" and "red line" for deep synthesis services, and maintain a healthy cyberspace ecosystem.

Effective as of December 1, 2023, the Ethics Review Measures aim to address the social and ethical challenges arising from the development of science and technology, boost innovation, and improve the regulatory frameworks and legal landscape for ethical review.

The measures set out detailed requirements for the procedures and standards of ethics reviews in the areas of science and technology. Activities that are subject to these reviews include scientific and technological activities (e.g., AI development) which involve humans or experimental animals, or may otherwise pose ethical challenges related to life and health, the environment, public order and sustainable development.

The European Union’s comprehensive AI Act was passed by the European Parliament on March 13, 2024. The Act entered into force on August 1, 2024, and will be effective from August 2, 2026, except for the specific provisions listed in Article 113.

The AI Act is designed to ensure AI developed and used in the EU is trustworthy, with safeguards to protect people's fundamental rights.

Specifically, the AI Act:

• addresses risks specifically created by AI applications
• prohibits AI practices that pose unacceptable risks
• determines a list of high-risk applications
• sets clear requirements for AI systems for high-risk applications
• defines specific obligations deployers and providers of high-risk AI applications
• requires a conformity assessment before a given AI system is put into service or placed on the market
• puts enforcement in place after a given AI system is placed into the market
• establishes a governance structure at European and national level

Additionally, the European Commission signed the Council of Europe Framework Convention on Artificial Intelligence in September 2024 on behalf of European Union. The Convention is the first legally binding international instrument on AI and is fully compatible with Union law in general, and the EU AI Act in particular, and provides for a common approach to ensure that activities within the lifecycle of AI systems are compatible with human rights, democracy and the rule of law, while enabling innovation and trust.

While there are currently no specific laws or legislation in India regarding AI regulation, various frameworks are being formulated to guide the regulation of AI, including:

• The National Strategy for Artificial Intelligence (June 2018), which aims to establish a strong basis for future regulation of AI in India.
• The Principles for Responsible AI (February 2021), which serve as India’s roadmap for the creation of an ethical, responsible AI ecosystem across sectors.
• The Operationalizing Principles for Responsible AI (August 2021), which emphasizes the need for regulatory and policy interventions, capacity building and incentivizing ethics by design with regards to AI.

While Japan currently has no law specifically directed to regulating AI, the Japanese government published new AI Guidelines for Business Version 1.0 in April 2024, which are not legally binding but are expected to support and induce voluntary efforts by developers, providers and business users of AI systems through compliance with generally recognized AI principles and following a risk-based approach.

Additionally, a government working group has proposed a regulatory law for AI, entitled the Basic Act on the Advancement of Responsible AI, which would adopt a hard law approach to regulate certain generative AI foundation models. Under the proposed AI Bill, the government would designate the AI systems and developers that are subject to regulation, impose obligations on them with respect to the vetting, operation, and output of the systems, and require periodic reports concerning such AI systems.

The Federal Council has tasked the Federal Department of the Environment, Transport, Energy and Communications (DETEC) with submitting a report by the end of 2024 to identify possible approaches to regulating AI in Switzerland. This analysis is meant to serve as a basis for the Federal Council to issue a specific mandate for drafting an AI regulation in 2025.

In collaboration with the University of Zurich, the Digital Society Initiative published a position paper in 2021, which outlines the approaches that should be taken to the legal coverage of algorithmic systems* in Switzerland, the issues that require particular attention, and how Switzerland should position itself in the context of European regulatory trends.

*the term “algorithmic systems” is used in the position paper instead of AI.

The U.K. government launched a 10-week consultation in December 2024 on “plans to give certainty to the creative industries and AI developers on how copyright material can be used to train AI models.” The goal is to help “drive growth across both sectors by ensuring protection and payment for rights holders and supporting AI developers to innovate responsibly.”

Additionally, before the change in government following 2024 U.K. general election, the government issued this white paper on its domestic AI regulation in March 2023.

Federal:
The U.S. lacks a unified AI regulation but has proposed or established numerous guidelines and frameworks to govern the AI sector on a federal level.

The Algorithmic Accountability Act of 2022 aims to hold organizations accountable for their use of algorithms and other automated systems that are involved in making critical decisions that affect the lives of individuals in the U.S. Among other requirements, the Act would mandate covered entities to conduct impact assessments of the automated systems they use and sell in accordance with regulations that would be set forth by the Federal Trade Commission. The Act is currently before the Senate Committee on Commerce, Science, and Transportation.

Currently before the U.S. House of Representatives, the No AI Fake Replicas & Unauthorized Duplications Act (“No AI Fraud Act”) is bipartisan legislation establishing safeguards to protect against generative AI abuses that stem from the unauthorized copying of a person’s individuality and result in deepfakes, voice clones, and non-consensual impersonations.

Currently before the U.S. Senate, the Nurture Originals, Foster Art, and Keep Entertainment Safe Act of 2024 (“NO FAKES Act”) legislation would create an enforceable new federal intellectual property right allowing victims of nonconsensual deepfakes and voice clones to have them quickly taken down and recover damages.

Note: this legislation was tabled before the change in government following the 2024 U.S. election. It is unclear if the new government will proceed with passing this legislation.

State:
State-wide legislation included below is specific to protecting performers and artists from the misuse of AI:

California
The state of California passed two AI bills on September 17, 2024, both of which will take effect January 1, 2025. Assembly Bill No. 1836 Use of likeness: digital replica protects against the unauthorized exploitation of digital replicas of deceased personalities, which will make it illegal to produce, distribute, or otherwise make available the digital replica of a deceased personality’s voice or likeness in an expressive audiovisual work or sound recording without prior consent. Assembly Bill No. 2602 Contracts against public policy: personal or professional services: digital replicas protects individuals (including performers) against vague, unfair, and unethical contractual terms that may, deceptively and without the performer’s full awareness, permit the unregulated production, use and distribution of digital replicas of their likeness.

Assembly Bill No. 2013 Generative artificial intelligence: training data transparency, which was passed on September 24, 2024, and will take effect January 1, 2026, requires AI developers to disclose whether they are training their models on copyrighted work.

Tennessee
The state of Tennessee, which was already just one of three states where name, photographs and likeness are considered a property right rather than a right of publicity, became the first state in the U.S. to enact legislation designed to protect songwriters, performers and other music industry professionals against the potential dangers of AI. The state’s “ELVIS Act” (the Ensuring Likeness, Voice, and Image Security Act), which came into effect July 1, 2024, ensures vocal likeness is now included in that list. The law also creates a new civil action where people can be held liable if they publish or perform an individual's voice without permission, as well as use a technology to produce an artist's name, photographs, voice or likeness without the proper authorization.