ARTIFICIAL intelligence is changing the way we create, communicate and consume information. One of the most concerning developments is the rise of deepfakes, videos, photographs and audio recordings created or manipulated using artificial intelligence to make it appear as though someone said or did something when they never did.
A person can now have their face placed onto another person's body, their voice replicated to say things they never said or their image used to promote a product they have never endorsed. What makes deepfakes particularly concerning is how realistic they can appear.
Although the legal implications of deepfakes are increasingly being discussed internationally, there has been comparatively little public discussion about the specific intersection between deepfakes, intellectual property and Zimbabwe's legal framework. This is a conversation Zimbabwe needs to have.
The question is no longer simply whether a deepfake is real or fake. The more difficult question is: who has the right to control the commercial use of your identity when artificial intelligence can reproduce it?
When your face and voice become commercially valuable
For many people, their name, face and voice are more than personal characteristics. They can have significant commercial value. Consider a popular Zimbabwean musician whose voice is replicated using artificial intelligence to create a song advertising a particular product. The musician never recorded the song, never approved the advertisement and never received payment. Or consider an entrepreneur whose face is digitally inserted into an investment advertisement, making it appear that they are recommending a financial product. The technology makes these scenarios increasingly possible. The problem is that traditional intellectual property law does not necessarily provide a straightforward answer.
Keep Reading
- Govt commissions CICs
- Letters: The life and struggle of students at the University of Zimbabwe
- ‘Hold security forces accountable for human rights abuse’
- Voters confront Zec
Copyright protects original literary and artistic works. Trademark law protects distinctive signs used in commerce. Performers' rights provide protection in relation to qualifying performances. But a person's face or voice is not automatically intellectual property in the same way as a song, book, invention or registered trademark. This creates an interesting legal gap.
A person may have considerable economic value attached to their identity, but the law may not always give them a simple and direct intellectual property right over that identity.
Zimbabwe is not without legal protection
It would be incorrect to suggest that Zimbabwe has no laws capable of dealing with deepfakes. Zimbabwe's Copyright and Neighbouring Rights Act provides protection for copyright works and also recognises certain rights of performers. These protections may become relevant where artificial intelligence is used to reproduce, manipulate or exploit an existing performance.
The Cyber and Data Protection Act also provide an important layer of protection. It regulates the processing of personal information and contains provisions dealing with cyber bullying and harassment, false data messages and the transmission of intimate images without consent. Depending on the circumstances, these laws may provide remedies to a person affected by a deepfake.
However, there is a distinction between having laws that may apply to certain aspects of deepfakes and having a legal framework specifically designed to address the challenges created by synthetic media. Deepfake technology is developing rapidly. The law, by comparison, often develops much more slowly.
Is copyright enough?
This is where intellectual property law becomes particularly interesting. Imagine that artificial intelligence creates a completely new video of a Zimbabwean musician singing a song that the musician has never performed. The musician's face and voice may have been replicated, but the video itself may be entirely new.
Who owns the copyright?
Can the musician claim copyright infringement simply because their face and voice have been used? The answer may not always be straightforward. Copyright generally protects original creative expression rather than a person's identity itself. This means that relying solely on copyright may not adequately address every situation involving deepfakes. The same problem can arise with a person's voice. A voice can be distinctive and commercially valuable, but that does not automatically mean that the individual owns a copyright in their voice. This is why the deepfake debate is increasingly moving beyond traditional copyright and into areas such as personality rights, performers' rights, privacy, data protection and the commercial appropriation of identity.
The commercial exploitation of identity
The commercial aspect of deepfakes deserves particular attention. A celebrity endorsement can be extremely valuable because consumers associate the celebrity with a particular product or brand. A musician may be paid to use their voice in an advertisement. An athlete may be paid to appear in a campaign. An influencer may build an entire business around their image and personality.
Artificial intelligence can potentially reproduce all of these things without the person's participation. Someone could create a convincing advertisement using an AI-generated version of a celebrity and make consumers believe that the celebrity supports the product. This is not simply a technological problem. It is an economic problem. The individual whose identity is being exploited may lose income, suffer reputational damage and have their commercial goodwill appropriated by someone else. The law, therefore, needs to consider whether existing protections sufficiently recognise the economic value of personal identity.
What about ordinary people?
It would also be a mistake to think that deepfakes only affect celebrities. The technology can affect ordinary people just as easily.
A fabricated audio recording could make it appear that an employee admitted to misconduct. A manipulated video could damage someone's reputation. A fake photograph could be circulated online. A person's image could be used to advertise a product or service without their consent. Deepfakes can also be used to create sexually explicit material involving people who never participated in its creation. The fact that the material is artificial does not mean that the harm is artificial. Once a fake image or video is circulated online, it can be copied, downloaded and redistributed almost immediately. Removing the original content does not necessarily remove the copies. For victims, the consequences can, therefore, be very real.
The African context
This issue is particularly important for Africa's growing digital economy. African musicians, actors, athletes, entrepreneurs and content creators are increasingly building valuable personal brands online. Social media has made it possible for individuals to reach international audiences without traditional intermediaries. This has created new opportunities for creators, but it has also created new vulnerabilities.
A Zimbabwean musician can have their voice reproduced by artificial intelligence and used anywhere in the world. A local entrepreneur can have their image incorporated into an advertisement that is distributed internationally. The geographical boundaries that traditionally made legal enforcement possibly become much less meaningful in the digital environment. This raises another important question.
If the person creating the deepfake is outside Zimbabwe, which country's laws apply? And even where Zimbabwean law provides a remedy, how practical is it for an individual to identify the person responsible and pursue them across borders? These are issues that policymakers cannot afford to ignore. The law needs to keep pace with technology. The answer is not necessarily to create laws that prohibit artificial intelligence or every form of synthetic media. AI has legitimate and beneficial uses.
It can assist filmmakers, musicians, advertisers, educators and other creators. Synthetic media can be used for entertainment, accessibility, education and artistic expression. The objective should, therefore, be to find a balance between encouraging innovation and protecting individuals.
Zimbabwe should consider whether existing laws provide sufficient protection against the unauthorised commercial use of a person's name, image and voice. There is also a need for greater clarity on the relationship among intellectual property, data protection, privacy, cybercrime and defamation when dealing with deepfakes. The possibility of recognising stronger personality or publicity rights could also be considered. Such rights could give individuals greater control over the commercial use of their name, image, voice and likeness. At the same time, any reform needs to protect legitimate activities such as satire, parody, criticism, journalism and artistic expression.
Who owns your digital identity?
The deepfake debate ultimately forces us to reconsider what we mean by ownership in the digital age. For generations, intellectual property law has focused largely on things that people create. Artificial intelligence presents another challenge. What happens when technology can recreate the person who created those things?
A person's face can become a brand. Their voice can become commercially valuable. Their image can generate income. Their identity can influence consumer decisions. Yet traditional intellectual property rights do not always provide a straightforward way of controlling that identity. Zimbabwe is, therefore,faced with an important policy question. Do our existing laws provide adequate protection for people whose identities are digitally reproduced without their consent? The answer may be that existing legislation provides some protection, but there are still gaps that need to be examined.
The conversation should begin now, before deepfake technology becomes so widespread that the law is forced to play catch up. Artificial intelligence is not going away. Neither are deepfakes. The challenge for Zimbabwe is to ensure that the law encourages technological innovation while ensuring that innovation does not come at the expense of individual rights. Perhaps the most important question for the future of intellectual property is no longer only “Who owns what we create?” It may soon also be: “Who controls who we become in the digital world?”