Liability of Internet Service Providers for Defamation in the US and Britain: Same Competing Interests, Different Responses

Michael Deturbide
Assistant Professor of Law
Dalhousie University , Halifax, Canada
[email protected]

Abstract

The liability of ISPs for defamation in the United States and Britain has been addressed by both the courts and legislatures in the respective countries. Early American decisions focused on distinguishing between ISPs that acted as publishers or distributors. Subsequent legislation in both jurisdictions has resulted in marked differences in the potential for legal liability of ISPs in America and Britain that supposedly reflect the inherent government policies of each country. These policies reflect a balancing of such interests as freedom of speech, personal reputation, and the promotion of electronic communication and commerce. The author argues that a liberal judicial interpretation of the relevant provisions of the U.S. Communications Decency Act of 1996 has exceeded the scope of government policy, whereas the U.K. Defamation Act 1996 does little to recognize the Internet as a unique communications medium.

This is a Refereed article published on 31 October 2000.

Citation: Deturbide M, 'Liability of Internet Service Providers for Defamation in the US and Britain: Same Competing Interests, Different Responses', 2000 (3) The Journal of Information, Law and Technology (JILT). <http://elj.warwick.ac.uk/jilt/00-3/deturbide.html>. New citation as at 1/1/04: <http://www2.warwick.ac.uk/fac/soc/law/elj/jilt/2000_3/deturbide/>

1. Introduction

The advent of the Internet has resulted in legislatures and courts around the world re-evaluating laws and policies on issues as diverse as taxation, privacy, and contract formation. The liability of the Internet Service Provider (ISP), the company that is the vehicle for the user's access to the Internet, and which brings information to the user from around the world, is potentially staggering if one applies to it long-established legal principles for issues such as distribution of pornography, breach of copyright, or misrepresentation. Defamation of character over the Internet is illustrative of the problem. ISP liability must reflect the need for law makers to balance the interests of its citizens who may be libelled because of postings accessed around the world, with the interests of society generally that use ISPs as conduits to this largely unfettered global communication medium.

Several factors make ISPs attractive defendants in defamation claims, many of which relate to the costs associated with litigation. For example, the author of a defamatory statement will often reside outside the jurisdiction of the plaintiff, whereas the ISP that carried the statement does business in the plaintiff's jurisdiction.[ 1 ] It might be difficult, time-consuming, or even impossible, to determine the actual author of the message. And even if the author can be identified, he or she may be judgement proof, whereas the ISP likely has 'deeper pockets'.

The liability of intermediaries for defamation has a long history in the common law.[ 2 ] 'Publishers', such as newspapers, which traditionally exerted editorial control over content, are generally liable for the defamatory statements that they publish. 'Distributors', such as bookstores or newsstands, exert very little if any editorial control, and have the benefit of the 'innocent disseminator' defence.[ 3 ] Innocent disseminators are protected from liability for defamation if they did not know of the libellous statement, there were no circumstances that ought to have led them to suppose it contained a libel, and they were not negligent in being ignorant of the libel.[ 4 ]

Many of the American cases that have considered the issue of ISP liability for defamation have focused on the possible analogies between print and broadcast media, and the Internet. A major difference in how ISPs conduct their business has been identified by some courts as a reason to deviate from standard principles of imposing liability for defamation on the ISP intermediary: ISPs, unlike their print and broadcast counterparts, generally do not impose any sort of editorial filter on content. The argument is then made that ISPs function more like the 'innocent disseminator', and should not be held liable for the defamatory statement carried on the service unless it was known, or the ISP ought to have known, of the existence of the defamatory statement.

The following paper will examine how courts in the United States and Great Britain have assessed this issue, analyse the legislative responses in the respective jurisdictions, critically evaluate the impact of the divergent directions taken in America and Britain on this issue, and assess the influence of these contrasting approaches in the rest of the common law world.

2 . The Early American Judicial Decisions

Two notable cases that dealt with the liability of ISPs that carried defamatory statements arose in the United States in the early '90s. These 'pre-legislation' cases demonstrated the willingness of the courts to apply existing principles to cases involving the Internet, and set the stage for legislative intervention to alter the application of these principles to ISPs.

2.1 Cubby, Inc. v. Compuserve, Inc.[ 5 ]

In 1991, the United States District Court, New York, considered the liability of Compuserve, an on-line service provider, for defamatory material carried on a special interest Journalism Forum. Subscribers to CompuServe's Information Service could gain access to the Forum, which included a newsletter entitled 'Rumorville USA'. Compuserve had contracted with Cameron Communications, Inc., to manage and review the contents of the Journalism Forum. Cameron Communications in turn contracted with Don Fitzpatrick Associates to provide the Rumorville newsletter. The contract between Cameron Communications and Don Fitzpatrick provided that Fitzpatrick accepted total responsibility for the contents of the newsletter. Compuserve had no opportunity to view the newsletter's contents before it was made available to Compuserve subscribers in the Journalism Forum.

The plaintiffs had developed a computer database which they intended to distribute electronically, in direct competition with Rumorville. The plaintiffs alleged that in response, Rumorville published false and defamatory statements about their service, and sued both Compuserve and Fitzpatrick. Compuserve moved for summary judgment that it acted as a distributor, not a publisher, of the statements, and therefore could not be held liable because it did not know and had no reason to know of the statements.

The Court quoted the general rule[ 6 ] that one who republishes defamatory statements is subject to liability, but that entities such as news vendors and libraries are not subject to liability if they neither know nor have reason to know the defamation.[ 7 ] The court reinforced the idea that distributors must receive such protection as a consequence of the constitutional guarantee of freedom of speech enshrined in the First Amendment , and made applicable to the states through the Fourteenth Amendment.[ 8 ] The court then held that Compuserve's product was 'in essence an electronic, for-profit library'.[ 9 ]

The court acknowledged that companies like Compuserve were 'at the forefront of the information technology revolution'.[ 10 ] As such, Compuserve made available the Rumorville newsletter instantaneously to subscribers, without any opportunity to exert any editorial control over the newsletter's contents. The court held that:

Thus, the court clearly classified Compuserve as a distributor, and applied the ordinary rules of distributor liability. Relying on a previous decision that held that computerized database services are entitled to the same protection, because of First Amendment guarantees, as more established modes of news distribution,[ 12 ] the court emphasized the new reality in information gathering and dissemination:

Accordingly, the burden of proving that Compuserve had knowledge or had reason to know of the defamatory statements shifted to the plaintiffs, who were unable to meet that burden. The court granted summary judgement in favor of Compuserve on the defamation claim.[ 14 ]

In deciding that Compuserve was a distributor, having no duty to monitor the information it distributes, the court did not assess what standard an ISP is required to meet in order to rely on the defence that it had no reason to know of a defamatory posting. For example, is it required to remove a posting if someone complains that it is defamatory? Given the thousands of postings that occur daily, one might argue that a required investigation of all such complaints would be overly burdensome. Nor did the court directly address the potential consequences of a service provider undertaking an editorial role. That issue was the central focus of the second pre-legislation U.S. case on ISP liability.

2.2 Stratton Oakmont, Inc. v. Prodigy Services Company [ 15 ]

In 1995, a judge of the New York Supreme Court issued a decision that left ISPs (at the time comforted somewhat by the previous Cubby decision) clamouring for legislative intervention for clarification (i.e. exoneration) of their liability for on-line postings. An unidentified party posted a defamatory comment on prodigy's bulletin board 'Money Talk' which, at the time, was one of the most widely read financial information bulletin boards in the United States. The posting stated that Stratton Oakmont, a securities investment firm, and its president had committed fraudulent acts in connection with an initial public offering, and further derogatory characterizations of the company and its employees. In order to advance their libel claim, Stratton Oakmont and its president sought partial summary judgment that Prodigy was a publisher of the allegedly defamatory statements, and that the Board Leader for the Money Talk bulletin board acted as Prodigy's agent for the purposes of the claim.

Prodigy relied on the Cubby decision in its defence. The court stressed that it was in full agreement with Cubby , and that computer bulletin boards were generally the equivalents of bookstores or libraries. However, according to the court, Prodigy's policies relating to postings on bulletin boards on the service altered the characterization of its service from 'distributor' to 'publisher'.

Prodigy advertised itself as being 'family oriented computer network', and in various national publications, it held itself out as exercising editorial control over the content of messages posted on its bulletin boards.[ 16 ] Prodigy used a software screening program that automatically removed offensive language, issued guidelines that indicated offensive posts would be removed, and hired bulletin board 'leaders' to enforce the guidelines. Thus, by making content decisions and establishing a system to monitor content, Prodigy exerted editorial control and therefore acted as a publisher. The court held:

The court underlined the fact that Prodigy had undertaken this decision to attract a certain market, and that in doing so, the company had to accept the concomitant legal consequences.[ 18 ]

The Court also determined that, on the facts, Prodigy had directed and controlled the actions of the bulletin board leader, who was responsible for monitoring and editing the Money Talk bulletin board. As such, the Court found that he had acted as Prodigy's agent.

Some commentators have argued that, considered together, the Cubby and Stratton Oakmont cases indicate that an ISP would be wise not to review content so as to avoid any risk of being considered a publisher.[ 19 ] This assertion is likely moot following the addition of Section 230 to Title II of the Communications Act of 1934 (discussed infra ). And while this assertion might be sustainable in American common law, an innocent disseminator under English law (who receives the protection of a 'distributor' highlighted in the above American decisions) must not only have no actual knowledge of a libellous statement, but must also exercise a reasonable standard of care.[ 20 ] The Court in Stratton Oakmont also took note of this possible presumption, which, according to the Court, ignores the market compensation that a service might receive for exercising control at the expense of possible liability.[ 21 ]

The Stratton Oakmont decision left many ISPs uncomfortable. ISPs that exerted an editorial role relating to violation of copyright or distribution of obscene material feared they could be liable for defamatory comments carried on the service. The U.S. Congress was urged to act to counter Stratton Oakmont and to clarify ISP liability generally in an era when the Internet was fuelling the heated high tech economy. The U.S. Congress response, encapsulated in communications decency legislation, and its subsequent interpretation by the courts provides an wide umbrella of protection for ISPs, and makes distinctions relating to degree or type of editorial content control irrelevant.

3. U.S. Communications Decency Act of 1996 [ 22 ]

The overriding concern reflected in the Communications Decency Act of 1996 (CDA) was, as reflected by the title of the statute, decency on the information highway. The Act was designed to extend the 'standards of decency' that protected users of other telecommunications systems to the Internet.[ 23 ] Several of the provisions of the law, which comprises Title V of the Telecommunications Act of 1996[ 24 ], were struck down in a highly publicized case[ 25 ] which stated that consumers of pornography must usually take active steps to access it.[ 26 ] However, the specific section of the Act relevant to liability of ISPs for on-line defamation (section 230) was not affected by this ruling.

Section 230, which provides protections to ISPs that take active steps to block or screen offensive material, was specifically enacted to counteract the ruling in Stratton .[ 27 ] The underlying principle of the section is that 'Good Samaritans' who undertake editorial duties to remove offensive content should not be penalized for their efforts by being treated as publishers, and hence be subject to liability for defamation or other causes of action. Section 230 (b) provides that it is the policy of the United States:

In furtherance of that policy, section 230(c)(1) states that no provider or user of an interactive computer service will be treated as the publisher or speaker of information provided by someone else, and section 230(c)(2) states that no provider or user of an interactive computer service can be held liable for voluntarily restricting access to or availability of objectionable material, or for making available the technical means to restrict access to such material.[ 29 ]

Whereas section 230(c)(2) would appear to address the concerns that followed the Stratton decision (arguably Prodigy, in undertaking an editorial function, would have been protected by this provision) section 230(c)(1) is very broad indeed. One need not be a 'Good Samaritan' editor to take advantage of section 230(c)(1) - any service provider is precluded from being treated as a publisher. Perhaps the only basis for liability that remained for an ISP under American law after the enactment of section 230 was distributor liability, as 'distributor' is not specified in the legislation. However, judicial interpretation on the scope of section 230(c)(1) has clearly indicated that the legislation, and the policies behind the legislation, required broad protection of ISPs from liability. Following these decisions, one might justifiably assert that, because of section 230, ISPs in the US are immune from liability for content carried on their services.[ 30 ]

4. Zeran v. America Online, Inc.[ 31 ]

In 1995, an unidentified person posted a message on a bulletin board hosted by the ISP America Online (AOL) advertising the sale of T-shirts that depicted tasteless and offensive slogans relating to the bombing of a federal building in Oklahoma City. Those who wished to purchase the shirts were invited to call the plaintiff's Seattle phone number. Zeran received a great number of phone calls, including death threats, from outraged callers. Because he relied on the phone number to operate his business from his home, he was reluctant to change the number. He contacted AOL to inform the ISP of his problem. He was assured by a company representative that the posting would be removed.

The original posting was removed[ 32 ]. However, several new postings appeared, advertising the sale of more offensive merchandise relating to the bombing, and listing Zeran's phone number. Zeran repeatedly called AOL, and was told that the account from which the messages were posted would be closed. However, similar postings continued to appear. An announcer for an Oklahoma City radio station became aware of the postings, and urged his listeners to phone Zeran's number to express their outrage. Zeran was inundated with calls and death threats from Oklahoma City residents. He required protective police surveillance. The calls did not abate until an Oklahoma City newspaper ran a story exposing the postings as a hoax.

Zeran sued the radio station that broadcast his phone number, and filed a separate action against AOL. He argued that once he had notified AOL of the anonymous postings, AOL had a duty to promptly remove the defamatory messages. The legal basis for the law suit was therefore the traditional test for liability of distributors, that is, the distributor had a duty to remove the material if it knew, or had reason to know, of its defamatory character.[ 33 ]

Both the district court and the Fourth Circuit Court of Appeals concluded that AOL was saved by s. 230 of the CDA. In so deciding, the Court of Appeal highlighted several policy reasons for the necessity of this statutory immunity. First, the Court stated that s. 230 was enacted 'to maintain the robust nature of Internet communication and, accordingly, to keep government interference in the medium to a minimum.'[ 34 ] The threat of tort-based lawsuits to free expression on the Internet required, according to the Court, such protection. Second, the Court recognized the virtual impossibility of ISPs screening millions of daily postings. Again, freedom of expression concerns were paramount. Without such protection, ISPs might severely restrict postings for fear of being sued. Third, the Court pointed to Congress' desire to promote self-regulation, and to remove the disincentives to self-regulation created by the Stratton Oakmont decision.

Zeran argued that s. 230 immunity applied to 'publishers'. The word 'Distributors' is not mentioned in the legislation, and therefore, according to Zeran, the principles of liability for distributors continued to apply. AOL, being a distributor of the information, could be held liable for the defamatory statements because they had the requisite knowledge and did not act to remove the postings in a timely manner.

The Court of Appeal rejected this distinction in the context of s. 230. The court concluded that although distributors are generally subject to a different standard of liability than publishers,

Although the Court merged the concepts of distributor and publisher liability for the purposes of s. 230, it has been argued that the Court expanded the interpretation of s. 230 beyond what Congress had intended, i.e., to overrule Prodigy , which did not deal with distributor liability.[ 36 ] Nevertheless, the broad ruling in Zeran , that 'by its plain language, s. 230 creates a federal immunity to any cause of action that would make service providers liable for information originating with a third-party user of the service'[ 37 ], indicates that ISPs are immune from all tort-based claims for the publication of third-party statements on the Internet.[ 38 ]

Clearly, then, the policies emphasized by the Court in Zeran have resulted in the Internet being treated, in the U.S., as a special medium of communication, deserving of special protection.[ 39 ]

5. Post Zeran Cases

The Zeran decision has had a wide reaching effect. Two subsequent cases, one dealing with defamation, the other with a negligence claim arising from the advertising of pornographic material, illustrate the influence of Zeran on the interpretation of s. 230.

5.1 Blumenthal v. Drudge [ 40 ]

In 1997, the Internet gossip columnist Matt Drudge stated in the Drudge Report that White House aid Sidney Blumenthal abused his spouse. Blumenthal sued both Drudge and AOL, which posted the Drudge Report on its service, for defamation.

The District Court for the District of Columbia held that, even though AOL received advance notice of the content of the Report, section 230 of the CDA and its interpretation in Zeran protected AOL from liability.[ 41 ] The court was clear that more traditional publishers, such as newspapers, could be held liable in similar circumstances for failing to check the accuracy of the statements.[ 42 ] The court emphasized the unique nature of the Internet, the importance of its growth and the imposition of measures that would prevent the stifling of its use, and the reflection of these policies by Congress in enacting s. 230.[ 43 ]

It is worth noting that although the claim against AOL was dismissed, the author of the allegedly defamatory statements was known, and his potential liability remained. Thus, the reasoning in Blumenthal produces a less severe outcome for the plaintiff than does the ruling in Zeran , where the author of the postings could not be identified. The reality of the Internet, however, is that the author of allegedly defamatory statements will often remain unknown, and the elimination of ISP liability leaves the plaintiff with no one to sue.[ 44 ]

5.2 Doe v. America Online, Inc.[ 45 ]

In an ironic twist, s. 230 of the CDA was held by the Florida courts to protect AOL from liability for failing to monitor content in this case, the marketing of child pornography. The mother of an eleven year old boy sued AOL for negligence and for violating state obscenity statutes for failing to monitor a chat room where an individual who had molested her son marketed videotapes and pictures of the assault. AOL also failed to remove the account of the child pornographer.[ 46 ]

Both the Florida state court and court of appeal relied on Zeran and s. 230 in dismissing the claims. The plaintiff raised the argument that even if AOL was protected as a publisher, it could still be held liable as a distributor. The Court of Appeal indicated that this issue had already been addressed, unsuccessfully, by the plaintiff in Zeran .[ 47 ] Thus, s. 230 of the CDA provided protection to AOL when it failed to remove the type of pornographic material which the CDA, by its very title, was supposed to tackle! It is doubtful that Congress intended that its 'Good Samaritan' provision, designed to protect ISPs who took active steps to monitor content, should shield those who irresponsibly or negligently failed to take any such steps.[ 48 ]

Although it does not deal with defamation, the Doe case illustrates the far-reaching consequences of s. 230 and its interpretation in Zeran. It may be that all tort-based claims against ISPs for material originating with a third party are barred because of section 230.[ 49 ] Although courts are sometimes uneasy with the outcome,[ 50 ] they appear to be constrained by the public policies on the Internet enunciated by Congress and the court in Zeran .

However, by passing such a broad provision as s. 230 to encourage ISPs to reduce offensive content, Congress may have unwittingly 'opened the door to irresponsibility by failing to provide any disincentive for Internet providers to engage in conduct, perhaps negligent conduct, that could lead to defamation.'[ 51 ] In effect, ISPs no longer have any incentive to be diligent or monitor content because of the immunity from liability provided by s. 230.[ 52 ]

One other recent US case merits some discussion, although the principles upon which it was decided ignore, strangely, s. 230 of the CDA.

5.3 Lunney v. Prodigy Services Company [ 53 ]

An unknown imposter posted two vulgar messages on a Prodigy bulletin board in the name of the plaintiff, and sent a threatening e-mail in Lunney's name to a third party. Lunney sued Prodigy for allowing the accounts to be opened in his name, and for defamation. Although the New York State Supreme Court initially denied Prodigy's applications for summary judgment, the Appellate Division granted summary judgment to Prodigy, holding that, inter alia , Prodigy was protected by the common law privilege that historically had been accorded telegraph and telephone companies.[ 54 ]

Lunney further appealed to the Court of Appeals of New York, which agreed with the earlier decision.[ 55 ] The court justified the common law privilege, as applied to the Internet, on policy grounds, not unlike the grounds articulated by the Court in Zeran for the necessity of s. 230 of the CDA. The Court stressed that the public would not be well served by requiring an ISP to screen millions of e-mail communications.[ 56 ] With respect to bulletin boards, the Court agreed with the Appellate Division that even if Prodigy exercised an editorial function, that did not alter its passive character considering the millions of postings that it did not screen.[ 57 ]

With respect to s. 230 of the CDA, both the Appellate Division and the Court of Appeals simply declined to address the issue as the case did not call for it.[ 58 ]

The Lunney decisions are unsatisfactory in many respects. The Courts had the Zeran decision before them, but gave it only passing reference. It is strange that the Courts were reluctant to consider s. 230 of the CDA, but rather engaged in a generic analysis of the common law privilege accorded to telegraph and telephone companies, whose application to the Internet contradicted at least one other recent defamation case based on similar facts ( Stratton Oakmont). In applying the common law test, the Court of Appeals simply stated that, on the facts, there was nothing to indicate that Prodigy was a publisher. It declined to state in what circumstances an electronic bulletin board operator would qualify as a publisher.[ 59 ]

The determination of whether e-mail and bulletin board communications are analogous to telegraph or telephone messages for the purposes of protecting the ISP from liability may be of limited analytical use given the broad wording of s. 230 of the CDA. Still, the comparison might be employed in future cases where the plaintiff attempts to narrow the interpretation or limit the application of s. 230.

6. ISP Liability in the UK

The law of defamation in the UK differs both in subtle and dramatic ways with the law of defamation in the US. Although an examination of these differences is beyond the scope of this paper, they will be highlighted as they arise in the context of ISP liability. The American decisions discussed supra have been considered by the High Court in the one decision dealing with ISP liability to receive judicial attention in the UK, and as the Court stated in that decision, the American decisions were 'educative'.[ 60 ]

6.1 Defamation Act 1996

In the context of ISP liability, the common law rules governing innocent disseminators have been supplanted by the wording of section 1 of the Defamation Act, 1996:

Section 1(1) is the statutory equivalent of the innocent disseminator defence. The defendant must establish all three requirements for the defence to succeed, unlike in American law where the plaintiff must prove that the defendant was not an innocent disseminator.[ 62 ]

Thus, because of section 1(3), an ISP will normally meet the requirement of s. 1(1), unless the ISP had effective control over the person who made the statement.[ 63 ]

Section 1(5) lists some factors to consider when determining the second requirement, taking reasonable care:

The UK legislation therefore adapts the traditional innocent disseminator defence to the on-line environment. It does not provide the carte blanche protection from liability that s. 230 of the American CDA does. An ISP, which by virtue of s. 1(3) is not an author, editor, or publisher; that takes reasonable care, having regard to the factors listed in s. 1(5); and does not know or have reason to believe that what he did caused or contributed to the publication of a defamatory statement, will be protected from liability for defamation.

If this is a logical interpretation of the legislation, it is interesting to contrast the outcome of the fact scenarios represented in the American cases, supra. For example, would Prodigy receive protection from liability in the UK if it had exercised the editorial control illustrated in Stratton Oakmont ? Perhaps not, since s. 1(3) deems that a person is not the author, editor or publisher of a statement, if he is only involved as the operator or provider of the communications system. Prodigy performed a greater role.[ 64 ]

Would AOL have been protected in the circumstances of Zeran ? Since s. 1(1)(c) requires that the ISP not know that it contributed to the publication, and s. 1(1)(b) requires that a standard of reasonable care be exercised, the facts in Zeran may have produced a different outcome under the UK legislation.

6.2 Godfrey v. Demon Internet Ltd.[ 65 ]

In 1999 a decision released by the High Court of Justice, Queen's bench Division, signalled that the U.K. Defamation Act 1996 would result in a markedly different approach to ISP liability for statements made by third parties than the positions taken by courts in the U.S. Godfrey was the first decision in the U.K. that dealt with defamation and the Internet. Although the parties ultimately settled their dispute, an earlier motion that was filed by Godfrey to strike out parts of the ISP's defence led to the judgment discussed herein.

Godfrey's action against Demon for defamation arose from a 1997 incident when an unknown person posted to the News group 'soc.culture.thai', a message purportedly originating with the Plaintiff that was, according to the court, 'squalid, obscene and defamatory'[ 66 ] of Godfrey. The posting invited replies by giving Godfrey's e-mail address.

The Defendant, Demon Internet, carried the News group and stored its contents for two weeks. The Plaintiff requested that the Defendant remove the offensive posting. It did not oblige, and the posting remained on its news server for two weeks. There was no dispute that Demon could have removed the posting at Godfrey's request.

Godfrey sued for defamation. In its defence Demon stated that it was not, at common law, a 'publisher' of the allegedly defamatory posting, and that even if it was, the defence of s. 1 of the Defamation Act 1996 was available.[ 67 ] The Plaintiff sought to strike out this part of Demon's defence pleadings, as disclosing no sustainable defence.

Godfrey succeeded in his application. In its decision, the Court reviewed the American authorities discussed supra , but cautioned that 'Care has to be taken before American cases are applied in English defamation cases'.[ 68 ] In reaching its judgment, the Court stressed the consequences of the Defamation Act, 1996, which did not adopt the American statutory approach.[ 69 ]

The Court agreed that Demon was not the publisher of the defamatory material within the meaning of the Act. Therefore, it satisfied the requirement of s. 1(1)(a)[ 70 ]. However, the Court found that the Defendant did not meet the requirements of s. 1(1)(b) and 1(1)(c)[ 71 ], which were also necessary to establish the defence. Demon knew of the posting but did not remove it. Therefore, according to the Court, the Defendant did not take reasonable care, and had reason to believe that it contributed to the publication of the defamatory statement.

The Defendant also relied on the Lunney [ 72 ] decision, which did not consider s. 230 of the CDA in its analysis. The English court found that under English common law Prodigy would have been considered to be a publisher in that case. Therefore, under common law principles or the application of s. 1 of the Defamation Act 1996, Demon had no defence.

Demon subsequently settled with Godfrey, which has inflamed the debate over freedom of expression on the Internet in the U.K. Demon hoped that the settlement would press the British government for recognition that ISPs should not be held liable for the millions of items that they carry daily.[ 73 ]

7. Balancing the Interests

The decisions on ISP liability emanating from the U.S. and the U.K. highlight the constant tension that surrounds defamation law: balancing the interests of free speech with the interests of personal reputation. The Internet adds other considerations, such as the practicality of reviewing millions of postings, which arguably renders historic distinctions between concepts such as 'publisher' and 'distributor' obsolete, or at least much less useful. Another consideration that is frequently voiced is the importance of the promotion of the use of the Internet, and the removal of disincentives that would limit its growth.[ 74 ] The commercial impact of the Internet on the individual state cannot be ignored by governments, and the UK must be cognizant of the European Union's proposed e-commerce Directive that contains provisions relating to ISP liability.[ 75 ] The relative importance of these various considerations has guided legislative authorities in different directions in the US and the UK.

S. 230 of the CDA certainly removes disincentives for ISPs to carry information provided by third parties. Yet, it is questionable whether Congress actually meant to absolve ISPs from all liability. There have been calls to rectify the 'mess' created by Zeran [ 76 ], which has actually provided a disincentive for ISPs to engage in the 'Good Samaritan' behaviour that is promoted in the CDA.

In the UK, s. 1 of the Defamation Act 1996 simply provides a:

Although this position is in keeping with established differences in the English law of defamation compared with its American counterpart, the Act does not deviate enough from the common law in recognizing the inherent problems created by the Internet. For example, an ISP is not considered to be a 'publisher' under s. 1(3)(e) of the Defamation Act 1996 if it exerts no effective control over statements transmitted over its service. This fulfils one of the requirements of the defence against defamation under s. 1(1). However, the other requirements of s. 1(1), which are traditional requirements for establishing the innocent disseminator defence, require the ISP to take reasonable care which, arguably, could result in exercising some control. It may be that the provisions of the Act create a classic 'Catch 22' state of affairs.

This is the type of contradiction that s. 230(c)(2) of the CDA was supposed to address, by providing that ISPs will not be held liable for voluntarily restricting access to or availability of objectionable material, or by making available the technical means to do so.[ 78 ] It is s. 230(c)(1) of the CDA[ 79 ] and its interpretation in Zeran and subsequent cases that derails the effect of s. 230(c)(2).

The later decisions from the US and the UK on the liability of ISPs may be of limited use in other common law jurisdictions, as they are predicated on the legislation of the respective jurisdictions. In Canada, for example, which adopts a position on protection of reputation that is more similar to that of the UK than the US,[ 80 ] defamation on the Internet has not thus far been distinguished in any way from defamation on any other medium. Absent special legislation, the courts will be faced with analyzing whether ISPs meet the traditional requirements of 'publishers' or 'distributors'. Therefore, pre-CDA cases from the U.S., such as Cubby [ 81 ] and Stratton Oakmont,[ 82 ] which rely on such distinctions in their analysis, may have to be revisited.[ 83 ]

The balancing of these diverse interests is not an easy task for governments or the courts. All are laudable policies, and the Internet unquestionably alters their dynamics. Yet the balance of these competing interests has so far produced unsatisfactory results on the issue of liability for ISPs in the US and the UK, the former for producing an enormous discrepancy between electronic and other media,[ 84 ] and the latter for not being aggressive enough in recognizing the uniqueness of the Internet. Legislation is required that clearly recognizes that ISPs, including those that undertake an editorial function, will not be liable prima facie for third party statements, but will lose such protection if reasonable measures are not taken to prevent defamatory statements from being posted around the world.

Footnotes

1 . Besides the fact that the plaintiff can sue in his or her particular jurisdiction, which provides cost savings and convenience, the multi-jurisdictional aspect of the Internet creates a wealth of other problems related to conflict of laws. See for example Braintech Inc. v. Kostiuk (1999), 171 D.L.R. (4th) 46, in which a resident of British Columbia Canada was sued for defamation in Texas. The British Columbia Court of Appeal refused to recognize the Texas judgment, as it was not demonstrated that the defendant's Internet posting had been read in Texas, and the defendant did not conduct business there. The court based its analysis on an earlier American decision, Zippo Manufacturing Company v. Zippo Dot Com, Inc., 952 F. Supp. 1119 (W.D. Pa. 1997).

2 . For an overview of the standards of liability of intermediaries at common law, see George Takach, Computer Law (Toronto: Irwin Law, 1998) at 395; for the U.S. see Annemarie Pantazis, ' Zeran v. America Online, Inc.: Insulating Internet Service Providers from Defamation Liability' (1999) 34 Wake Forest Law Review 531 at 537-542.

3 . The defence has been codified in various common law jurisdictions, eg. s. 1 of the U.K.Defamation Act 1996, discussed infra , and in the common law provinces of Canada ( eg. Nova Scotia's Defamation Act , R.S.N.S. 1989, c. 122, s. 16(5)).

4 . Takach, supra note 2 at 396. The American test is similar, in that distributors are not liable if they neither knew or had reason to know of the defamation. See Restatement (Second) of Torts s. 581 (1977).

5 . 776 F. Supp. 135 (S.D.N.Y. 1991).

6 . Restatement (Second) of Torts s. 578 (1977).

7 . Lerman v. Chuckleberry Publishing, Inc., 521 F.Supp. 228 (S.D.N.Y. 1981) at 235.

8 . Smith v. California , 361 U.S. 147 (1959).

9 . Supra , note 5 at 140.

10 . Supra , note 5 at 140.

11 . Supra , note 5 at 140.

12 . Daniel v. Dow Jones & Co., 520 N.Y.S. 2d 334 (N.Y. Civ. Ct. 1987) at 340.

13 . Supra note 5 at 140

14 . The court also dismissed the claim that Compuserve could be held vicariously liable for the Rumorville statements, holding that neither Cameron Communications nor Fitzpatrick were agents of Compuserve.

15 . [1995] WL 323710 (N.Y. Sup. Ct. 1995).

16 . Although Prodigy did indicate that its policies had evolved over time, and that the last such statement regarding editorial control was made over a year before the allegedly libellous statements were posted.

17 . Supra note 15, at p. 12 of 16.

18 . The Court took note of the fact that, at the time of the decision, the Communications Decency Act of 1995 was before Congress, which would ultimately pre-empt the issues addressed in the decision.

19 . Davidson, S.J., Podkopacz, J.S. and Dechery, L.J., 'The Law of Cyberspace: Liability of Information Service Providers', December 1999 Journal of Internet Law, online: < http://www.gcwf.com/articles/journal/jil_dec99_1.html >

20 . Takach, supra note 4 at 404.

21 . Supra note 15 at p. 12 of 16.

22 . 47 U.S.C.A. § 230.

23 . 141 Cong. Rec. S1953, as cited by David R. Sheridan, ' Zeran v. AOL and the Effect of Section 230 of the Communications Decency Act upon Liability for Defamation on the Internet' (1997) 61 Albany Law Review 147 at 159.

24 . 47 U.S.C.A. § 223.

25 . ACLU v. Reno , 929 F. Supp. 824 (E.D. Pa. 1996)

26 . Ibid. at 844.

27 . See David Weiner, 'Negligent Publication on Electronic Bulletin Boards: Is There Any Liability Left After Zeran ?' (1999) 39 Santa Clara L. Rev. 905 at 929.

28 . 47 U.S.C.A. § 230(b)(4).

29 . 47 U.S.C.A. § 230(c)(1), (2).

30 . It has been argued that if Congress meant to include protection from liability for distributors, they would have used 'distributor' terminology. Stratton did not impose distributor liability, and section 230 was specifically intended to override Stratton . Moreover, since the case law has historically differentiated between 'publishers' and 'distributors', it would be reasonable to assume that the word 'publisher' does not subsume 'distributor'. See David R. Sheridan, ' Zeran v. AOL and the Effect of Section 230 of the Communications Decency Act Upon Liability for Defamation on the Internet' (1997) 61 Alb. L. Rev. 147 at 159-162.

31 . 958 F. Supp. 1124 (E.D. Va. 1997), aff'd. 129 F. 3d 327 (4th Cir. 1997).

32 . Although the parties disputed the time it took AOL to remove the original posting.

33 . Restatement (Second) of Torts, s. 581 (1977).

34 . Supra note 31 at 330.

35 . Supra note 31 at 332.

36 . See A. Pantazis, ' Zeran v. America Online, Inc.: Insulating Internet Service Providers from Defamation Liability' (1999) 34 Wake Forest Law Review 531 at 549.

37 . Supra note 31 at 330.

38 . See Wiener, supra note 27, where the author examines the effect of the Zeran ruling on situations involving liability for the posting of misleading or inaccurate information.

39 . This assertion is demonstrated by the possibility that a print newspaper may be sued for publishing a defamatory letter to the editor, but would be protected from liability if the letter was published in the on-line version of the newspaper. See Sheridan, supra note 30 at 149.

40 . 992 F. Supp. 44 (D.D.C. 1998).

41 . Ibid at 53.

42 . Ibid at 49.

43 . Ibid at 48-53.

44 . See Andrew J. Slitt, 'The Anonymous Publisher: Defamation on the Internet after Reno v. American Civil Liberties Union and Zeran v. America Online ' (1998), 31 Connecticut Law Review 389 at 413-414.

45 . 718 So. 2d 385 (Fla. Dist. Ct. App. 1998).

46 . Ibid at 386.

47 . Ibid at 389.

48 . See Davidson et al , supra note 19.

49 . See Weiner, supra note 27.

50 . For example, the court in Blumenthal concluded that AOL had taken advantage of all the benefits of the CDA without accepting any of the burdens, by making rumours and gossip instantly accessible and then claiming immunity when a defamation suit inevitably resulted. See Blumenthal v. Drudge, supra note 40 at 52.

51 . Slitt, supra note 44 at 415.

52 . As a further irony, public entities such as libraries, which are not covered by s. 230, and that take steps to exercise editorial control, risk being considered 'publishers' and subject to liability. It is perverse that corporate ISPs are protected from defamation suits, whereas public entities that attempt to filter offensive content receive no such protection. See Michael H. Spencer, 'Defamatory E-Mail and Employer Liability: Why Razing Zeran v. America Online Is a Good Thing' (2000) 6 Richmond Journal of Law and Technology 25.

53 . 683 N.Y.S. 2d 557 (A.D. 2 Dept. 1998), aff'd. 723 N.E. 2d 539 (N.Y. Ct. App. 1999).

54 . 683 N.Y.S. 2d 557 (A.D. 2 Dept. 1998), at 562.

55 . 723 N.E. 2d 539 (N.Y. Ct. App. 1999).

56 . Ibid. at 542.

57 . The Appellate Division stressed that Prodigy had abandoned its policy of extensive editorial control prior to the events in question. Therefore, the Stratton Oakmont decision did not apply. The Court also criticized Stratton Oakmont for penalizing conduct that ought to be encouraged. Supra note 54 at 562.

58 . Supra note 53 at 563 and 543 respectively.

59 . Supra note 55 at 542.

60 . Godfrey v. Demon Internet Ltd., [1999] E.W.J. No. 1226 (Q.B.D.), para. 1.

61 . Defamation Act 1996, c. 31.

62 . Supra note 60.

63 . Quare whether AOL would have been a publisher under 1(3) of the UK legislation in the Blumenthal v. Drudge case, supra note 40.

64 . This approach taken by the UK legislation has been criticized by some commentators. On the one hand, the ISP must demonstrate reasonable care. However, if the ISP exercises editorial control, it risks losing its status as a mere provider under s. 1(3). See Lilian Edwards, 'Law and the Internet: Regulating Cyberspace, Defamation and the Internet: Name Calling in Cyberspace' (1997), online < http://www.law.ed.ac.uk/c10_main.htm >

65 . [1999] E.W.J. No. 1226 (Q.B.D.)

66 . Ibid. at para. 12.

67 . Ibid. at para. 2.

68 . Ibid. at para. 1.

69 . Ibid. at para. 45.

70 . Supra note 61.

71 . Supra note 61.

72 . Supra note 53.

73 . B.B.C. News online: <http://news.bbc.co.uk/hi/english/sci/tech/newsid_965000/695596.stm> [no longer valid link]

74 . See, for example, s. 230(b) of the CDA, supra note 28 on the policy of the United States in enacting s. 230.

75 . Article 14(1) states that Member States must provide in their legislation that ISPs will not be held liable for information stored at the request of a recipient of the service, provided that the ISP is not aware of facts or circumstances from which illegal activity is apparent, or the ISP, upon obtaining such knowledge, acts expeditiously to remove or to disable access to the information.

76 . See Michael H. Spencer, 'Defamatory E-Mail and Employer Liability: Why Razing Zeran v. America On-Line Is a Good Thing' (2000) 6 Richmond J. Law and Tech. 25.

77 . Per Lord Mackay in discussion of the 1996 Defamation Bill, as reproduced in Godfrey , supra note 65 at para. 2.

78 . Supra note 29.

79 . Supra note 29.

80 . See Hill v. Church of Scientology , [1995] 2 S.C.R. 1130, in which the Supreme Court of Canada stressed the importance of reputation to Canadians.

81 . Supra note 5.

82 . Supra note 15.

83 . Although even these cases may be of limited use in Canada, where the common law innocent disseminator test would not protect an ISP in circumstances where the ISP was negligent in carrying a defamatory message.

84 . See Sarah Beckett Boehm, 'A Brave New World of Free Speech: Should Interactive Computer Service Providers Be Held Liable for the Material They Disseminate' (1998) 5 Richmond Journal of Law and Tech. 7.