Generative AI: A Fun Sandbox or a Potential Legal Nightmare in the Making?

Entrepreneurs, developers, and individuals involved in AI technology and creative fields, take heed: the Next Big Thing you build could be a potential legal nightmare. If not navigated carefully, the hidden legal risks associated with generative AI could lead to serious consequences, knowingly or unknowingly.

Last week, I came across a post on X, a response from a famous YouTuber, Marques Brownlee, to an AI entrepreneur who made a recommendation aggregator using his photograph and domain booked using his YouTube brand MKBHD (askmkbhd.com)

The urgency of going from Idea to Market is becoming more challenging to contain for people building AI tools and utilities because everyone wants to be noticed, showcase their talent, capture opportunities, seek funding and more.

Here, the developer has a good use case for building a recommendation engine for a specific niche, and the question arises: If he had made such a tool without using the celebrity name or reference, would it have garnered the attention it got after Marques responded to his post?

I captured interactions of some X users from this post to get an idea of how the mindset of entrepreneurs is wired to think that nothing is off limits and everything is permitted. Celebrities, actors, singers, and artists from creative fields have started getting serious about the threats they face with the growing use of generative AI, which, even done for fun, can cause damage to their public image or upfront violation of privacy, IP, and copyright laws.

Let us break down the crux of this case:

1)??The utility developer claims he did this for non-commercial purposes and fun but did not seek permission from the person whose name, image, and brand identity (MKBHD) he is using.

2)??Adding a disclaimer that ‘We are not affiliated with MKBHD’ does not matter.

3)??The developer claims that he is using publicly available data, which could be debated, and it is more like a grey zone when it comes to the use of AI.

4)??He has trained the AI program to pull together content created by Marques, mostly his YouTube videos.

When I started reading comments from other users, I realized how ignorant we can be about basic copyright and privacy laws. Our sense of entitlement has grown after the LLMs arrived on the scene. Data misuse possibilities have grown multifold, and the associated complexities will continue to grow.

Some people suggested that Marques not overreact and that it was harmless; another commented that this was public data and that he is a public figure; MKBHD should instead encourage and acquire his solution.

This brazen attitude leads to lawsuits, and one court verdict will create more unnecessary barriers for genuine innovators. Web scraping may not be illegal now, but cyber and privacy laws will eventually catch up to define clearly what is permissible and what is off-limits.

There is a reason why media and publishing companies are now taking to court to hold big tech accountable for using content they created for their generative AI programs for various reasons. Even if Marques Brownlee has his videos posted on YouTube, between both entities, they hold the rights to published content, and they can issue a takedown of any such AI program that aggregates results from scraped content without explicit permission. This developer's rookie mistake is using MKBHD’s brand identity and photograph without permission, a strict no-no that does not even come under AI ambiguity. He cleaned up the web page, but the domain remains askmbhd.com.

Marques is a content creator who owns the rights to his content, whether behind a paywall or publicly available; he could monetize his content to build his own AI-powered recommendation utility, probably partnering with brands via commercial deals.

My advice to young and emerging entrepreneurs is to refrain from assuming stuff and go on to ship AI solutions that garner negative attention and probable legal trouble. The laws related to AI use cases are evolving, and I do not expect developers to keep a tab on legal stuff or know everything in and out. If you have been a developer long enough, you would have undergone data protection and privacy training. You should try to stick to the premise of existing copyright and privacy guidelines from those trainings, and you should be fine.