What counts as
AI-generated content?

Any creative work produced with the help of AI models and tools, ranging from written text and images to music, videos, and computer code, is considered AI-generated content. So whether you’re using AI to draft up a work email, help you write lyrics to a new song, or create images for your website, you’re creating AI-generated content. 

Most AI-generated content begins with a prompt, like create a mystery novel plot or design a new logo for our company. Some outputs get heavily edited — you might change the color, scale, texture, and shape of an AI-generated logo to better suit your needs. Some AI-generated content gets released into the real world exactly as it is. Think: blog posts that read like a robot wrote them. 

However, most content falls somewhere in the middle, and this raises a critical legal, ethical, and philosophical question. In many countries, current copyright guidelines suggest that ownership can be claimed if a human makes “meaningful changes” to AI-generated work but how do you determine meaningful human contribution?

Not only is this standard vague and subjective, it also brings up a host of new questions: 

• Is choosing the best version of several AI outputs “meaningful”? 

• What about rewriting a few sentences or adjusting the lighting in an AI-generated image?

• How about editing text for flow and grammar?

This ambiguity sits at the heart of the authorship and ownership debate. How do we know when we’ve adjusted the AI-generated content enough to make it our own?

Who should own
AI-generated content?

To understand how the public feels about this issue, a recent study conducted by Toloka and LMU Munich asked the big question: who owns the content? Over 300 participants from 49 countries responded — and opinions varied widely:

• 50% believe full authorship belongs to the user.

• 18% believe full ownership belongs to the user, even without authorship.

• 19% are uncertain.

• 13% advocate for shared or platform ownership.

The study surveyed AI Trainers, which adds an interesting layer to the results. Although these individuals train AI models and tend to have a deeper understanding of how they work, their views on authorship are complex and divided. 

Curious, Mindrift conducted our own survey. We asked over 100 AI Trainers at Mindrift and here’s what we found:

• 55% use AI tools for writing and creating content every day.

• The top 3 uses for AI are expanding on topics, brainstorming, and grammar and writing flow checks.

• 59% believe they own the content created in collaboration with AI. 

When we asked how they determine authorship, the majority (57%) believe the human is the author, regardless of the level of AI contribution. Asked to elaborate on this belief, one survey respondent explains:

The response of the AI depends on the quality of my prompts, so they are my creation in the end.

Another respondent minimized the role of AI in content generation, explaining:

I see AI as a creative assistant, a technical tool, like a pen or a text editor. It supports the process, but the ideas, structure, and final decisions are still mine. The authorship remains clearly human, guided by intent, context, and critical thinking.

At the same time, a large portion of respondents (22%) believe that authorship is shared, while a smaller segment (8%) think it’s impossible to determine. One survey respondent shared their concerns:

I sometimes worry about how much of the final content is influenced by the AI and whether that could be questioned in professional or academic settings. Clear guidelines would help.

Another respondent brought up a different side of the debate — is the AI model stealing and reusing content?

I think if generative AI is used for writing content instead of you, authorship is outright unethical. You are basically stealing a mix of a bunch of other authors’ generated content (and nowadays it's even questionable whether the content that AI uses is human made or has been made by AI).

Another respondent points out that while AI is a helpful tool, the content it creates will always need a human touch:

AI is a powerful assistant, but content created with it should always be reviewed, refined, and owned by a human, ensuring integrity, accuracy, and originality.

The takeaway? While the majority still see the user as the creator, there's no consensus — and a growing awareness that AI is changing not just how we create, but how we determine creativity itself.

A legal take on the debate (from around the world)

Countries around the world are taking different approaches to authorship and ownership of AI-generated content by defining what type of content is applicable for copyright laws. 

• United States: Only works with meaningful human creativity qualify for copyright. Simply entering a prompt isn't enough for legal protection.
  
  

• United Kingdom: Copyright can apply to AI-generated works, with ownership given to the person who made the necessary arrangements for the creation.
  
  

• China: AI-generated content may be protected if it shows clear human input and originality, aligning with traditional copyright principles.
  
  

• European Union: Copyright remains tied to human authorship, though some member states are exploring potential reforms for AI-specific scenarios.
  
  

• Japan: AI-generated works are generally not protected by copyright, reflecting a preference for open innovation and public domain use.

An ethical gray area and more questions

So who deserves the credit — tool users, developers, or no one? Does the credit go to the original creators of the content that was used to train the AI models? And how do we make these decisions?

Opinions are expectedly divided. Even experts across different industries don’t quite agree. Take creative writing, for example. Recently, the Society of Authors protested Meta’s creation of LibGen, an AI training database that contains over 7.5 million books. 

“A book can take a year or longer to write. Meta has stolen books so that their AI can reproduce creative content, potentially putting these same authors out of business,” explained SoA chair Vanessa Fox O’Loughlin. 

Meanwhile, a growing number of authors are blatantly using AI models to write their books, creating a nearly finished product to publish. A recent controversy brought this to the public’s attention when author Lena McDonald accidentally forgot to remove a ChatGPT prompt from her published novel. 

When it comes to AI-generated art, Dr. Alfredo Cramerotti and Auronda Scalera, two industry experts, share a more positive view, saying:

“Many critics argue that A.I. will undermine human labor and lead to creativity devaluation, which in turn will create repetitive or formulaic art. But if we look harder, history dictates otherwise. Photography did not destroy painting; it pushed artists toward modernism. The same is true with the print and text-based outputs.”

The authorship and ownership debate doesn’t just end at who owns this content — it goes much further. This ethical gray area creates more questions that no one has the answers to. 

“If a company uses AI-generated content for messaging, and over time, another firm produces an identical message, can the former claim ownership of its AI-generated content?” asks Deepa Nagraj, Global Head of Communications & Sparkle Innovation Ecosystem at Mphasis, in a recent Forbes article. This question brings up a thought-provoking point: how many times can we reuse AI’s output until we inevitably end up “copying” each other?