Photo: Andrew Feinberg, CC BY 2.0.

Last week, I had the opportunity to talk to Mozelle Thompson at an event organised by QUT IPKCE and the IIA. Mozelle was a US Federal Trade Commissioner, and is now a legal adviser to Facebook.

Mozelle had some interesting things to say about Facebook's privacy policies, including that he had recently spoken to a convention of Australian police officers and reinforced Facebook's pledge to only cooperate with law enforcement if it's demands were backed by valid legal procedure (ie., warrants and judicial oversight).

”[Facebook is] not here to provide people who want to spy the process for a fishing expedition.”

Interestingly, he also mentioned that Facebook would not comply with authorities if it believed the local laws to be too onerous:

”[…] if we think that that legal process is overbroad or inappropriate we will not enforce it.”

Mozelle also had some statistics about Facebook's adoption and growth. Without going into the details, Facebook is becoming enormous, everywhere. This raises some interesting questions. It is slowly becoming difficult to organise and participate in events if you're not a member of Facebook. Indeed, a large proportion of the people at Mozelle's talk heard about it only through Facebook. Both social and professional networking appears to be migrating to Facebook and (perhaps to a lesser extent) similar platforms. Anecdotally, it appears that it is not uncommon for a friendship group to organise themselves almost exclusively through Facebook. This means that people are becoming increasingly reliant on Facebook and other proprietary platforms for the organisation of their social life, and, to my mind, this raises the question of what responsibilities do Facebook and such other platforms have to their users?

The first thing to point to is Facebook's Terms of Use, which could charitably be described as oppressive. Significantly, they contain a clause which allows unilateral termination:

The Company may terminate your membership, delete your profile and any content or information that you have posted on the Site or through any Platform Application and/or prohibit you from using or accessing the Service or the Site or any Platform Application (or any portion, aspect or feature of the Service or the Site or any Platform Application) for any reason, or no reason, at any time in its sole discretion, with or without notice[.]

So, I asked Mozelle whether there were any applicable limits to Facebook's discretion in the way it treats its users. His answer was a very emphatic 'no'. Mozelle highlighted that it doesn't make good business sense to treat your customers poorly, and that the market provides essential safeguards for the interests of users. The problem with this argument is fairly simply stated: (a) there are few alternatives to facebook; (b) network effects inhibit exit and impose barriers to entry to new players; and © even if the market were efficient, it reduces important issues of rights and interests to a market rhetoric and provides little to no protection for the interests of minority groups.

When I pressed Mozelle about non-market limitations, he forcefully rejected the suggestion that Facebook's discretion could be limited. By differentiating Facebook from public utilities, Mozelle argued that the public (government) had no right to interfere in the way in which Facebook was run. Unlike public utilities, Facebook receives no public funding and operates in a competitive market. I asked for clarification on this point, because the market certainly doesn't seem competitive. Mozelle, who was a Federal antitrust lawyer, answered that you couldn't definte the market as narrowly as 'a market for social networking websites', and, accordingly, Facebook didn't have market power.

Competition law aside, there is a significant problem with the false dichotomy presented by Mozelle Thompson. It is simply not true that either an entity is a public utility, in which case it is regulated, or it is a private proprietary corporation, in which case it is not. We impose limits on the behaviour of private entities all the time. There is no reason that we cannot alter the boundaries of private property and the apportionment of liability and responsibility in any given case, and the proper location of these boundaries is exactly the discussion we need to be having.

I pointed Mozelle to the example of Sara Andrews, who was threatened with banning from World of Warcraft by Blizzard after she advertised for a LGBT-friendly guild. Mozelle distinguished World of Warcraft from Facebook based upon the subscription fee which WoW gamers pay. The gist of the argument was that by paying $15/mo, WoW subscribers had a right to complain to Blizzard about the way they are treated. Facebook users, on the other hand, pay nothing, and therefore have no such rights.

This is very dangerous thinking. Firstly, Facebook users, in aggregate, provide almost all of the value of the Facebook company. The technical platform represents some intiial investment, but it is the social network which provides the real driving force (and advertising revenue). If we focus only on subscription models, we are able to say that there is no consideration paid by Facebook users, and hence no contractual remedies. On the other hand, if we recognise the value that participants provide, this assumption may no longer hold.

More importantly, I believe that a valid contractual relationship is not the only source of liability that platforms like Facebook may be exposed to. There are any number of non-contractual arguments which could be raised, including, most significantly, negligence, estoppel, and unjust enrichment. It may be that Facebook owes its users a duty of care not to arbitrarily or maliciously remove them, for example. Alternatively, it may be that Facebook's oppressive Terms of Use are not adequately reflected in the internal community norms, and Facebook may be estopped from enforcing those terms as written in a particular case.

We need to stop talking in terms of clear dichotomies between private and public spaces. These private networks are providing functions which were public in nature when we drew the boundaries we know – which explains why there are constitutionally protected remedies against the State when it prevents you from associating with your social network in public. In no way does this fact preclude us from determining the appropriate level of responsibility that proprietors will owe to individuals in the future.

It may well be that we will decide not to impose liability on facebook for arbitrarily or maliciously ejecting its customers, but this result is by no means certain. By presenting these issues as a clear dichotomy between private and public, we are ignoring the malleability of legal rules and forestalling a proper debate on the rights and responsibilities of actors in our networked society. We are also ignoring the very real harms that individuals may suffer at the hands of platform owners like Facebook, and it is certainly time open up this debate. In this debate, the reification of property-based arguments will only slow us down.

Where the proprietor of a virtual community generally fails to enforce the rules, could it be estopped from doing so in any particular instance?

Despite a clear contractual right to terminate, a provider may be estopped from terminating in circumstances where it would be unconscionable to do so.¹⁾ In order to prevent the provider from terminating, a participant would have to show that the provider had represented that it would not terminate, that the participant relied on that representation to his or her detriment, and that it would be unjust or inequitable for the provider to terminate in those circumstances.

The representation that the provider would not rely on the right to terminate does not need to be explicit, but it must be unequivocal.²⁾ The representation does not need to be made to a particular person, but can be made to a class of people.³⁾ Where breaches of a particular rule are widespread, a long-standing failure to enforce the rule could conceivably be construed as a representation that the provider will not enforce the rule in future. However, such non-enforcement could also be construed as not making any representation as to the future.⁴⁾ The requirement that the representation be unambiguous does not mean that “it cannot possibly be open to different constructions, but that it must be such as will be reasonably understood in a particular sense by the person to whom it is addressed.”⁵⁾ Whether a representation has been made is a question of fact, and its existence “must be decided on ordinary common law principles of construction and of what is reasonable, without fine distinctions or technicalities.”⁶⁾ In practice, while possible, it may be quite difficult for a participant to establish that the provider represented that it would not enforce a rule in future or against any particular person.

If a representation can be shown, the participant must also be able to show that he or she reasonably relied on that representation. While showing reliance may be straightforward – in that the participant would not have engaged in conduct that technically broke the rules if he or she did not believe that the rule would not be enforced – showing that the reliance was reasonable may be more difficult. In Galaxidis v Galaxidis, Tobias JA (with whom the other members of the NSW Court of Appeal agreed) held that

the representation is sufficiently clear and unambiguous if it is reasonable for the representee to have interpreted the representation in a particular way being a meaning which it is clearly capable of bearing and upon which it is reasonable for the representee to rely.⁷⁾

Brennan J, in Walton Stores v Maher, held that it was “essential to the existence of an equity created by estoppel that the party who induces the adoption of the assumption or expectation knows or intends that the party who adopts it will act or abstain from acting in reliance on the assumption or expectation”.⁸⁾ Again, whether it is reasonable for a participant to rely on a representation that the provider will not enforce a strict contractual right will depend on the circumstances.

It may be difficult to establish in many cases, but it would certainly be open for a judge to find that a platform owner is estopped from terminating a particular participant's access to the virtual community where it takes no action against others who have habitually broken the same rule. If an estoppel can be established, it is important to consider that estoppel does require that that the representation or promise be fulfilled, but instead only provides a remedy for the detriment suffered as a result of reliance upon the representation.⁹⁾ For this reason, an estoppel, unlike an election, is not permanent – if the detriment to the relying party can be cured, the provider will once again be entitled to exercise its rights. For practical purposes, this means that given sufficient warning, a provider may be able to begin to enforce rules which it had largely ignored in the past.

There is a lot of flexibility in the doctrine of estoppel, and a significant normative question arises as to whether it ought to apply in any given case.¹⁰⁾ In the most extreme cases, it will almost certainly be effective as a brake on the ability of platform owners to rely on strict contractual rights which it has encouraged participants to believe would not be enforced. Its application in other circumstances, however, will depend in a large part on the discretion of the court as to how the alleged representation is interpreted and how reasonable the court believes the reliance on that representation to be. Like the other ways in which strict contractual rights can be read down, I believe that these considerations will depend particularly on the importance which the court attaches to the interests of the participant at issue.

Most EULA and ToS documents now require the participant to 'accept' the document by checking a box or clicking a button before continuing. If this is accepted to be equivalent to signing the contract, then the terms of the documents will generally be incorporated regardless of whether the participant has actually read them or not (absent any misleading conduct).¹⁾ However, if clicking an 'I agree' button is not treated as equivalent to signing a contract, and is instead considered more analogous to the ticket cases (incorporation by notice), terms will only be incorporated where the drafting party can show that he or she has taken reasonable steps to bring the clause to the attention of the other party.²⁾ What is reasonable, of course, depends on the circumstances and on the clause itself.

Most of the ticket cases concern exclusion clauses. It appears likely, however, that it is the unusualness, severity, or unreasonableness of the clause which is important, rather than the strict type of clause. Lord Justice Denning (as he then was), in J Spurling Ltd v Bradshaw, noted, obiter, that

the more unreasonable a clause is, the greater the notice which must be given of it. Some clauses which I have seen would need to be printed in red ink on the face of the document with a red hand pointing to it before the notice could be held to be sufficient.³⁾

Lord Denning MR had an opportunity to develop this reasoning in Thornton v Shoe Lane Parking Ltd, where the Queen's Bench was asked to determine whether an exclusion provision referenced on a pillar upon entering a car park and again in small print on a ticket dispensed by an automatic parking machine was validly incorporated in the contract. His Lordship held that it was not, because the ticket was issued after the contract was formed;⁴⁾ however, he noted, obiter, that if this were not the case, then the customer would be “bound by the exempting condition if he knows that the ticket is issued subject to it; or, if the company did what was reasonably sufficient to give him notice of it.”⁵⁾ Lord Denning appeared to recognise that the more onerous the provision, the greater the requisite notice must be – in the instant case, the limitation was
so wide and so destructive of rights that the court should not hold any man bound by it unless it is drawn to his attention in the most explicit way.⁶⁾

In the same case, Megaw LJ applied a test of whether or not unusual terms had been “fairly brought before the notice of the accepting party”.⁷⁾ His Lordship held that

at least where the particular condition relied on involves a sort of restriction that is not shown to be usual in that class of contract, a defendant must show that his intention to attach an unusual condition of that particular nature was fairly brought to the notice of the other party. How much is required as being, in the words of Mellish LJ, “reasonably sufficient to give the plaintiff notice of the condition,” depends upon the nature of the restrictive condition.⁸⁾

This principle was explicitly considered by Brennan J in the High Court in Oceanic Sun Line Special Shipping Co Inc v Fay, where His Honour said that

where an exemption clause is contained in a ticket or other document intended by the carrier to contain the terms of carriage, yet the other party is not in fact aware when the contract is made that an exemption clause is intended to be a term of the contract, the carrier cannot rely on that clause unless, at the time of the contract, the carrier had done all that was reasonably necessary to bring the exemption clause to the passenger's notice⁹⁾

In Baltic Shipping Co v Dillon (The Mikhail Lermontov) (1991) 22 NSWLR 1, Kirby P (as he then was) considered that there was a “responsibility to bring unusual conditions at least to the notice” of passengers of a cruise ship – it was the unusualness of the limitation clauses, combined with the failure to alert passengers that they ought to make their own arrangements for insurance, which meant that simply stating that the contract was subject to terms and conditions was not sufficient notice.¹⁰⁾ Similarly, it was the “subject matter and content of the relevant limitation clauses”¹¹⁾ which “significantly limit[ed] the appellants common law liability”¹²⁾ that determined the requisite level of notice for Gleeson CJ (as he then was) in the same case. Chief Justice Gleeson concluded that “it is the fact, and extent, rather than the precise mechanics, of the limitation that are of primary importance.”¹³⁾

In Interfoto Picture Library Ltd v Stiletto Visual Programmes Ltd, the Queen's Bench did not accept that the reasoning of Shoe Lane Parking was limited to exclusion clauses, Dillon LJ holding that

what their Lordships said was said by way of interpretation and application of the general statement of the law by Mellish L.J. in Parker v South Eastern Railway Co, 2 CPD 416, 423-424 and the logic of it is applicable to any particularly onerous clause in a printed set of conditions of the one contracting party which would not be generally known to the other party.¹⁴⁾

Lord Justice Dillon extended the decision in Shoe Lane Parking to the general principle that

where a condition is particularly onerous or unusual the party seeking to enforce it must show that that condition, or an unusual condition of that particular nature, was fairly brought to the notice of the other party.¹⁵⁾

Lord Justice Bingham reached the same conclusion, holding that

what would be good notice of one condition would not be notice of another. The reason is that the more outlandish the clause the greater the notice which the other party, if he is to be bound must in all fairness be given.¹⁶⁾ at “a very high and exorbitant rate”,¹⁷⁾ which resulted in “an inordinate liability.”¹⁸⁾ Indeed, the both judges in this case apparently would have held that the offending clause was void as a penalty clause, but the argument was not raised at first instance or upon appeal.¹⁹⁾ In the circumstances, the Queen's Bench had no difficulty substituting a quantum meruit for the subject matter of the contract.²⁰⁾ Interfoto has not, however, been greatly extended in the two decades since it was decided.

The ticket cases are interesting for our purposes. They show a line of authority which suggests that where a person does not read a contract (and is not reasonably required to read the contract) then any surprising terms must be reasonably brought to their attention before they will be bound. These cases suggest that it must be possible to contract without reading the whole terms, and that it is the responsibility of the drafter to make surprising or unusual terms stand out. They are likely to be negated, however, where the contract has been signed.²¹⁾

The interesting point to draw from this line of cases is that if, for some reason, the ticket cases can be extended to cover click-wrap contracts (ie., the terms are sought to be incorporated by notice rather than by assent), then there is support for the proposition that surprising and unusual terms should be specifically brought to the attention of the participant, without the need to imply a duty of good faith. This in turn is interesting, because it particularly accords with the Karl Llewellyn's theory of Blanket Assent.