How AI is quietly reshaping the battle between users and Big Tech

I’m waging a silent rebellion against Amazon. The internet giant owns Goodreads, the leading service for tracking books, but treats it like aging plumbing. So I used AI to build my own version, breaking the terms of service without a second thought. Take that, Amazon.

I’m not alone. Every day brings fresh examples of people using AI to unofficially remake software: free Photoshop clones, ad-stripped smart TVs, hacked printers on unsupported devices. It’s all happening so fast that lawyers haven’t caught up yet, though in egregious copyright cases, they eventually will.

The key to my rebellion? I built the app to avoid mass scraping of Goodreads data, which would be easy to detect. Instead, it mimics human behavior and flies under the radar.

The AI Agent Problem

This connects to a bigger battle: AI agents. Companies like OpenAI and Meta are launching tools that can shop or book appointments for you. Amazon and JPMorgan Chase are already fighting back, blocking them aggressively. The irony? Meta blocks protest accounts on Instagram after detecting activity that “resembled automated bot traffic rather than ordinary human browsing.”

All of this raises an ancient question: where do your rights end and theirs begin? We’ve seen this play out repeatedly in tech history, and past battles offer clues about what’s coming.

Consider web browsers. In the late 1990s, Microsoft dominated with Internet Explorer and blocked competitors like Opera from accessing MSN. Browsers responded by disguising themselves as other browsers. Today, the right to use any browser won. Individual choice triumphed, though smartphone apps have since shifted power back to companies.

Then came ad blockers. Once your browser fetches a website’s HTML, should you display all the ads? Around 30 percent of users now block them. Companies and blockers are locked in an arms race, with no clear winner.

Learning from History

The grandfather of all internet copyright battles was Napster. Illegal music sharing seemed unstoppable until Spotify arrived. Legal alternatives killed piracy almost overnight, though now most people don’t own the music they listen to.

Internet users won the browser wars. Ad blocking reached stalemate. Streaming music mostly benefitted corporations. So what does history suggest about AI?

The crucial difference is decentralization. I can build my own Goodreads app, and so can you, each slightly different. Blocking them all without restricting humans would be nearly impossible. Sure, Amazon could theoretically pressure OpenAI or Anthropic to restrict such software, but open-weight AI models that anyone can download make this a losing game.

Most likely, we’re heading toward a three-part future. Anything you do privately on your computer will be nearly impossible to stop. Release it online, and expect lawyers. We’ll probably see prompts shared widely that let AI do the work in private, instructions that no lawyer can touch. Individual choice will prevail, but the choices will be microscopically granular.

For software that touches the wider internet, including AI agents, expect two tiers. Companies like Meta and Amazon will eventually reach agreements allowing certain agents to function. A “personal agent protocol” already exists through partnerships with Shopify and Stripe. For commercial users, corporate rules apply.

But why use commercial agents? People with basic technical skills are already building independent alternatives. As AI improves at writing code, barriers drop lower. Companies will have to accept this, much like they accepted ad blockers. Soon, a substantial portion of users who look human will actually be custom-made bots doing things solely for their creators’ benefit.

These bots will be ungovernable, for better and worse. The question isn’t whether we can stop them, but whether we’re ready for a world where we can’t tell humans from machines anymore.

Source: New Scientist