From nonexistent attractions to fake book lists: should we stop AI from generating content for us?

AI can do everything. Even resurrecting a dead singer to create another song.

At its core, AI was created to help us with repetitive, dull, and annoying tasks. However, somehow, it has started targeting mediums that people enjoy creating most: music, videos, writing, and all other types of content.

To be more specific, not that long ago, some newspaper readers were surprised after reading a “books-to-read” list that doesn’t actually exist. The book list was generated by AI and ended up in multiple physical publications without an editor or proofreader batting an eyelid.

This situation has already raised some red flags about AI users getting carried away, using the tool to do all their work for them rather than just using it as an aid.

And when it comes to physical books, written content is threatened not only by faux or non-existent information but also by the fact that it's been used to train AI and destroyed in the process.

AI’s recent threats to the music industry

Written content isn’t the only medium that “suffers” from AI’s impact – it has entered the music industry, too.

The Velvet Sundown, now often referred to as a synthetic band, raised suspicion after gaining a significant number of monthly listeners (at one point, it reached 1.5 million), which were accused of being generated by AI.

The band tried to deny the claims. However, it also didn’t provide much proof that real people were behind the music. After releasing two albums, it was finally admitted that the whole thing, the band, its music, shared images, and its backstory were generated by AI.

Soon after, Spotify had to take down a song by country artist Blaze Foley, as it was generated by AI. What gave it away? Perhaps the first clue is that the artist died in 1989, making it a bit of a stretch to be releasing new songs in 2025.

Considering how fast artificial intelligence is taking over the content we consume, is now the time to draw some conclusions and try to stop the spread?

Which type of content is mostly generated by AI?

The chances of stumbling upon AI while reading a text are higher than you might think.

“Available tooling largely dictates the types of content that are generated by AI. Text generation is the most ubiquitous given the early availability of GenAI systems like ChatGPT, Claude, Grok, and others that focused on text generation,” notes Mike Price, chief technology officer at ZeroFox, a digital risk protection company.

Next follows image generation, which has grown in popularity in the last few years. The expert also shared that while audio and video generation remain less popular, this could change in the future thanks to new tools and features.

Now that AI content generation has grown in popularity, does this mean that we have become better at detecting it?

In general, there are a few ways such content can be detected: either through “human analysis,” which means being able to see repetitive patterns that AI carries, or by employing the same AI to tell us if AI made the content we’re seeing.

Certain nuances can also make it even harder to detect whether a text is generated by AI.

“Content that is entirely generated by AI from a single prompt, without human editing or guidance, can be more easily detected due to a lack of nuance, depth, or a creative spark. That said, detection isn't always straightforward,” notes Paula Hijosa from Space & Time marketing agency.

Is AI taking over “physical” content?

Have you ever found it difficult to distinguish between the real and online worlds? For many, it’s becoming harder to draw a distinct line between the two.

The current situation with AI doesn’t help, as many experts agree that it is already impacting physical content, such as books, TV shows, music, and more.

“This is not inherently bad when it comes to fiction, but for things like travel books and non-fiction books, the hallucinatory tendencies of today’s LLMs is concerning with respect to the need to provide reliable information to consumers,” noted Hany Farid, co-founder of cybersecurity company GetReal.

For one elderly couple, this concern became a reality when they decided to take a 186-mile trip to see a tourist attraction that wasn’t real but rather a figment of AI’s imagination.