There's a 100% Chance AI Agents Are Ruining the Internet
AI-powered agents flooding websites, inboxes, and customer support channels are seen as accelerating the decline of the open, human-centered internet, intensifying spam, friction (CAPTCHAs, bot checks), and paywalled “walled gardens.” Commenters debate whether solutions should be legal (regulating automated spam), technical (proof-of-work, micropayments, stronger bot detection), or social (stigmatizing AI-mediated contact), but many doubt any approach can be enforced globally or against bad actors. A recurring theme is disillusionment: people report retreating to offline life, older media, and smaller, gated communities as the mainstream web fills with low-quality AI-generated content and defensive countermeasures.
Legal and technical responses to AI-driven spam
- Many argue the core problem is mass automation of communication (email, calls, form fills), not “AI” per se.
- Proposals: require agents to self-identify, use legal restrictions on automated spam, micropayments, or proof-of-work.
- Pushback: hard to enforce across borders and bad actors; proof-of-work and micropayments have environmental and usability downsides.
- Some fear that turning to governments or central infra providers (e.g., “one big company”) to solve this just replaces one unaccountable power with another.
Erosion of the open internet and rise of walled gardens
- Several see AI agents as accelerating the “tragedy of the commons”: open sites get flooded by bots, so operators lock content behind paywalls, CAPTCHAs, aggressive bot checks, or logins.
- Cloudflare and similar protections often degrade human access (e.g., blocked from downloading appliance manuals), while bots still get through or pay to pass.
- Captchas are increasingly annoying, sometimes inaccessible, and reportedly no longer effective against smarter agents or cheap captcha‑solving services.
Shifts in personal internet use
- Multiple commenters report using the internet less for discovery and socializing and more for narrow tasks (e.g., code, specific information).
- People describe moving to books, older content (pre‑2022/2024), in‑person meetups, phone calls, and local events.
- There’s nostalgia for a “golden era” of a nerd‑dominated web (variously dated from pre‑1993 to pre‑2010), and pessimism that it will return.
AI agents in customer interactions and commerce
- Strong aversion to AI answering phones or chat for small businesses; seen as worse than big‑company bots and destroying a key “human” advantage.
- Examples of customers refusing contractors or services that use voice bots; some leave negative reviews when they encounter AI support or AI food photos.
- Counterpoint: in some contexts (e.g., dealing with warranty support, finding doctors who accept insurance), people would welcome agents doing tedious calls.
- Debate over whether “minor” AI shortcuts (menu photos, marketing details) signal deeper neglect of quality and hygiene, with industry professionals insisting they often do.
Content quality, spam, and detection
- Complaints about “AI slop” in papers, manga, YouTube videos, shorts, and social media; some now avoid recent content unless heavily vouched for.
- AI‑driven inbound spam emails are getting more targeted and “personalized,” sometimes referencing obscure résumé details.
- Attempts to spot AI by stylistic tics (e.g., em‑dash spacing) are seen as unreliable and unfair; style guides vary.
Responsibility, norms, and outlook
- Some stress that humans deploying agents are to blame, not the tools, and call for strong social stigma against using agents to contact real people.
- Others are skeptical that stigmas work online; they may simply drive stealthier agents.
- A minority embrace an “accelerationist” view: let AI finish breaking the already‑enshittified internet so people spend more time offline.
- A few argue that despite the mess, LLMs usefully amplify individual creators, and focus should include positive possibilities and new “Internet 2” models, even if current human‑oriented protocols don’t scale to agent traffic.