OpenAI's Sora Is Raising Big Legal Questions

ICYMI, OpenAI recently launched Sora, its latest tool for creating realistic, minute long videos from simple text prompts.

It’s both unprecedented and terrifying.

Footage by OpenAI showcased several short videos, including a Shiba Inu wearing a beret and black turtleneck, two surfers riding a giant tidal wave in a historical hall and more – all made within minutes. There’s no question the imagery is impressive, but it also raises questions about possible legal challenges, including copyright infringement.

When Sora was launched, OpenAI did not divulge how its AI model was trained. However, they do admit that it learns using both publicly available videos and videos licensed from copyright holders.

Now, that’s the big problem.

How Copyright Affects AI

Because Sora is also trained using publicly available videos, the owners of these videos can raise copyright violations and intellectual property issues at any time.

This can lead to AI companies facing plenty of copyright claims.

Copyright is the automatic protection of the expression of original ideas. It is the legal right that gives copyright owners the right to control certain activities with their works. These activities include the reproduction and re-use of material, such as:

●????? publication

●????? performance

●????? adaptation

●????? and communicating the work to be public (for example, making it available online)

?

With Sora having the ability to mix and match elements from various sources to produce the required video from the prompts, there’s no stopping the original owners of the videos from saying they did not allow their work to be used this way. In fact, there have been plenty of instances where authors, artists and news organisations have sued AI companies, accusing these businesses of copying their works in order to train their AI models. They also claim they did not give permission or receive any payment for the use of their work in AI output.

?

This is despite AI companies arguing that publicly available online images, text and video should be considered "fair use"? under copyright law.

?

Ethical Concerns Surrounding AI

?

There is also the question of ethics.

?

Are the videos produced by artificial intelligence accurate, ethical or biased? This is something that needs to be examined closely and continuously.

?

While Sora is currently available for testing to a limited audience – a small group of academics, outside researchers and select filmmakers – it has the potential to be misused for fake and misleading content once it becomes widely accessible.

?

Imagine this: lifelike but fake videos going viral on social media with the intent to destroy public trust in political leaders, institutions and the media. It’s downright scary.

?

In fact, it’s already starting.

?

Every day, we are getting closer to a time when it will be difficult to distinguish fake from real. Add lifelike, AI-generated audio to the mix, and you’ll be hard-pressed to separate genuine videos from those that are not.

?

These convincing deepfakes raise serious ethical and privacy issues, which underscore the need for close scrutiny and regulation.

?

Not to mention, an AI model capable of generating realistic videos poses a threat to several industries, including the film, entertainment, visual and even stock footage industry. Around the world, workers have complained that they have lost their jobs as businesses shift to using AI tools.

?

For its part, OpenAI is acknowledging all potential issues by delaying Sora’s public launch and performing “red team” exercises. These exercises involve experts who try to break through the AI model’s safeguards to identify potential scenarios where it can be misused. According to the company, these people are experts in areas like misinformation, hateful content, and bias.

?

OpenAI also has strict parameters for the AI tool, especially for applications related to sex and violence. They have also added watermarks to videos produced by Sora, tagging them as “generated by AI.” However, they acknowledge that users may have the capacity to remove these watermarks.

?

The Role of the Law in AI

?

On the legal side, lawyers must work harder to fight the darker side of artificial intelligence. It’s going to be increasingly difficult, for example, to charge someone for a crime based on video footage that is so real but yet so fake, and it is up to humans to determine that.

?

Legislators, solicitors, and barristers will face a real challenge going forward, and we’ll have to adapt quickly because this is the world we are moving towards.

?

Sora is a revolutionary tool, but it also highlights the problems in current laws and policies that have yet to catch up with rapidly evolving technology.

?

AI companies have argued that creating useful and helpful AI models is impossible without accessing copyrighted material. Are we then stifling technological advancement because of copyright?

?

Like other emerging technologies, AI presents big opportunities as well as challenges. This is precisely why lawyers and lawmakers face a tough battle as it progresses. They need to strike the right balance – giving room for tech platforms to innovate while also protecting access, rights and ethics. How they handle it will shape AI’s future.

?

What’s your take on this controversial topic? Is AI changing the world for the better, or is it a threat?