Antispam Bills: Worse Than Spam?

(Wired News) While no one has sympathy for the devils that fill inboxes with promises of lower mortgages and larger members, not everyone is supporting the new movement to banish spammers from the Internet.  Some online advocates worry that heavy-handed antispam measures, such as centralized blacklists and charging for delivery, will destroy e-mail.  Electronic Frontier Foundation’s head counsel Cindy Cohn, for instance, argues that antispam crusaders are forgetting the Internet’s first principle — information flows freely from end to end. Cohn fears that the Internet’s openness will be collateral damage in the war against unwanted e-mail. Cohn says her organization’s position on spam blocking can be boiled down to a simple proposition: “All nonspam e-mail should be delivered.” It’s an information age take on the Hippocratic oath, which requires doctors to first do no harm.

“It’s not the job of an ISP to block e-mail,” added Cohn. “E-mail isn’t a toy anymore. If I don’t get an e-mailed notice from the federal district court mailing list, it’s malpractice.”

Even some who sell antispam software to companies say that ISPs shouldn’t be blocking mail.

“Blocking e-mails is folly,” said Brian Gillette, whose company sells an enterprise-level, antispam appliance called trimMail Inbox. “If I’m an ISP and I stop a $150,000 equipment sale because I decided it was spam, I’m in for a lawsuit.”

Marv Johnson, legislative counsel for the American Civil Liberties Union, worries that the ability to speak anonymously on the Internet is being put at risk by federal antispam legislation.

Howard Beale of the Federal Trade Commission told House members at a recent hearing on proposed antispam legislation that “spam threatens to destroy e-mail.”

Several of the bills currently under consideration would make it illegal to mask a sender’s identity or forge routing information, both of which are tricks used by spammers to avoid the ire of those who receive their e-mails. But it’s also a tactic used by dissidents in countries with repressive governments who want to communicate with like-minded individuals.

“Many of these bills criminalize a message header that isn’t accurate,” said Johnson. “That’s not fraud. If you send out messages through an anonymizer, then you could get sent to jail.”

Cohn concurs, arguing that the bills criminalize the behavior of people — such as closeted gay teens or government whistle-blowers — who have legitimate reasons for speaking anonymously on the Internet.

EFF already has been a victim of overbroad spam filters. Its newsletter, which has more than 30,000 subscribers, has been bounced by aggressive keyword filters. In one case, its message was blocked because it contained the word “rape,” used when talking about EFF’s advocacy on behalf of an online group, Stop Prisoner Rape.

When the EFF asked around, it found that other noncommercial bulk mailers, such as listservs, were running into problems, too.

For example, AOL blocked e-mails from one of EFF’s clients, MoveOn.org, an online, liberal political action group which saw its membership swell to more than 2 million during the antiwar movement.

“MoveOn.org, one of EFF’s clients, has problems all the time, but MoveOn.org is now big enough to be on whitelists,” said Cohn. “I’m more concerned about the next MoveOn.”

Challenge-and-response systems pose particular problems for newsletters and listservs. These systems try to cut down on fraudulent e-mail by not delivering a message until the sender replies to a confirmation e-mail sent by the intended recipient’s ISP or e-mail host.

“Declan McCullagh of Politech and Dave Farber of Interesting-People can’t do 100 challenge-responses a day,” said Cohn. “That, as a solution, doesn’t scale.”

It would be wrong to call Cohn soft on spam. While in private practice she sued a spammer and won a court injunction and $60,000. And her employer uses antispam technology on its own servers.

The difference, according to Cohn, is that the SpamAssassin software EFF uses doesn’t block spam, it simply rates each e-mail. Staffers then set up their e-mail clients to separate messages into different inboxes. This keeps the main e-mail boxes free of spam, but allows individuals to check the spam folder occasionally to see if a legitimate e-mail was incorrectly tagged as junk.

Many in the technology industry think that only better technology can stop the spam deluge.

“The only people who can stop spammers are other technologists,” said trimMail’s Gillette.

The most promising new approach is better filters that use Bayesian algorithms to tag spam automatically and move it into a spam folder. The algorithms look at the body and header of an e-mail and judge from past experience whether an incoming message is junk. Users then train the algorithm, by moving misclassified e-mail from one e-mail folder to another.

Paul Graham, who many credit for applying Bayesian filtering to the spam problem, is ecstatic at the power of the new filters.

“I don’t need blacklists,” said Graham. “My own software is better than I am at deciding what is spam and what is not.”

Several open-source and commercial products, such as SpamBayes and Spam Bully, already use Bayesian filtering.

The ACLU’s Johnson hopes the new technology will head off the worst of the antispam legislation.

“Why do we want to start imposing a different world for the Internet than we have in the real world?” asked Johnson.

“Let the marketplace handle spam,” he said. “When Congress wants to show they are doing something about an issue, they often screw it up.”