The Flood Is Already Here

This isn't a warning about the future. It's a report on what's happening right now.

Late last month, dozens of near-identical AI-generated songs named some variation of "Angel Above Me" or "Run Run River" went viral on Spotify and TikTok simultaneously, according to The Atlantic. Some hit No. 1 on iTunes in Germany and Austria. Millions of streams. Real money.

The original song? "Angels Above Me" by the reggae band Stick Figure, released in 2019. The actual co-writers aren't always credited on the AI copies. So the people who created the thing that made the AI copies possible are getting cut out of the profits.

That's theft with extra steps.

106,000 Songs Per Day

According to data from analytics firm Luminate, 106,000 songs — AI-generated and otherwise — were uploaded to streaming platforms every single day in 2025. The platforms' spam-filtering systems can't keep up. DIY distribution services will push virtually any track to Spotify, YouTube, or Apple Music for a small fee, with minimal vetting for copyright violations.

AI music has charted before. But song generators are now fast enough and good enough to flood the zone before human reviewers can react. The Atlantic's reporter generated a convincing AI clone of Kendrick Lamar's voice in 30 seconds. An AI version of an Elliott Smith song with the original's exact lyrics took just as long.

This is a legal and ethical vacuum, not a technical one.

It's Not Just Music

The same dynamic is hitting the written word.

Steve Rosenbaum published a book called The Future of Truth, about how AI distorts reality. Irony died on the day that book shipped. The New York Times reported the book contained over half a dozen made-up or misattributed quotes. Rosenbaum admitted in a statement to including "a handful" of "improperly attributed or synthetic" quotes.

Wired ran the book's full text through Pangram, which the outlet calls the current gold standard in AI detection. The result: 53 percent AI-generated, with an additional 9 percent flagged as likely AI-assisted. That's the majority of a book about truth being apparently written by a tool that hallucinates facts.

When Wired's head of research asked Rosenbaum directly how he used AI, he responded that he used it for "source discovery, brainstorming, structural feedback, and language refinement" — while insisting the ideas and authorship were his. He holds a master's degree in "truth" from New York University.

The Transparency Con

Here's the line everyone is using to dodge accountability: AI is just a "writing tool," no different from spell-check.

That framing is dishonest. Spell-check doesn't invent quotes. Spell-check doesn't write your paragraphs. Spell-check doesn't replace your voice with a statistically averaged simulacrum of human prose.

According to The Atlantic, submissions to literary publications now arrive "perfectly clean, without a stray comma; uniform in length, with evenly paced paragraphs and a distinctive tone that was simultaneously breezy and grandiose." Authors who six months ago would apologize for an AI-assisted paragraph now openly defend it.

The competitive pressure is real. Journalism, academia, grant writing, YouTube — all are brutally crowded. People are using AI because it gives them an edge in volume and polish. But understandable and honest are not the same thing.

What Mainstream Coverage Is Getting Wrong

Most coverage frames this as a technology story — AI detection tools, platform safeguards, legal gray areas. The actual issue is accountability, or its complete absence.

When a musician uploads an AI clone of Stick Figure's song without crediting the original writers, that's stealing. When an author publishes AI-generated quotes as real ones in a nonfiction book, that's fraud.

Platforms like Spotify, Amazon, and Apple are making money off this content and have every financial incentive to move slowly on enforcement. Publishers greenlit The Future of Truth without catching fabricated quotes. Literary editors are accepting AI-written submissions they can't reliably detect.

Nobody is being fired. Nobody is being sued. Nobody is losing their book deal.

What the Law Actually Says — And Doesn't

Covers require licenses. Samples require clearance. When Ariana Grande interpolated a tune from The Sound of Music on "7 Rings," she gave up 90 percent of songwriting royalties to Rodgers and Hammerstein, according to The Atlantic. That's the system working.

AI remixes exist in a legal gray area that courts haven't fully defined yet. Platforms' terms of service are patchwork. Copyright law was written before anyone imagined a machine that could clone a voice in 30 seconds for free.

Congress hasn't acted. The platforms are slow-walking enforcement. And the creators getting ripped off don't have the legal budgets of the tech companies profiting from the infrastructure that makes this possible.

What This Means

If you're streaming music, you may be unknowingly enriching AI operators instead of the artist you think you're supporting. If you're reading a nonfiction book, some of those quotes may be fabricated. If you're getting a text message, an email, or even a dating app message — there's a growing chance a human didn't write it.

The person who crashed into your car and sent you a perfectly grammatical explanation of the accident an hour later? That wasn't his voice. The mechanic quoting you for repairs in polished paragraphs? Same deal.

This isn't about hating technology. It's about a basic standard: if you didn't make it, don't take credit for it. That used to be common sense.