The internet is essentially anonymous. Online activity cannot be traced or used against you.
Internet activity leaves extensive digital traces - IP addresses, browser fingerprints, cookies, ISP logs, and server logs. This was true from the internet's inception. The expectation of anonymity was always a misconception.
What changed?
The New Yorker cartoon ran on July 5, 1993. In it, a dog at a computer terminal turns to another dog and explains: "On the Internet, nobody knows you're a dog." Peter Steiner's panel captured something real about how the early internet felt, a space of chosen pseudonyms and screen names, freed from the social categories that governed everyday life. For a generation coming of age online in the late 1990s, this felt like a design feature.
It was never accurate. Every internet connection routes through a numeric address, an IP address, assigned by an Internet Service Provider. ISPs log which account connected from which address and when. Web servers log every request made to them. By the time a student in 1998 sent an email, browsed a website, or posted to a bulletin board, they had left behind a paper trail distributed across their ISP's records, the servers they'd contacted, and any routers those packets had passed through. Anonymity required active technical effort; the default state of internet use was identifiability.
The legal and corporate mechanisms to exploit that record were still developing in the late 1990s, which may be why the anonymity myth persisted. But cases were already establishing the principle. A German court convicted a CompuServe executive in 1998 over content hosted on its servers. Courts in the United States were beginning to force ISPs to identify subscribers accused of copyright infringement. The architecture of traceability had been there from the beginning; the social and legal machinery to exploit it simply followed within a few years.
The real cost of the anonymity myth was not abstract. Students who understood online activity as untraceable wrote things, shared things, and built digital histories they later found unexpectedly durable and discoverable, in college admissions, in employment background checks, in legal proceedings. The internet was not a consequence-free parallel world. It was a logging system with a social interface. Teaching that distinction in 1998 curricula would have required grasping something that most adults who made policy had not yet understood either.