AOIR Regulating Virtual Worlds Panel, and My Notes on Investment Expectations in Virtual Worlds

By Eric Goldman

Last week at AOIR’s annual meeting (AOIR 8.0) in Vancouver, Greg Lastowka, James Grimmelmann, Tyler Ochoa and I presented on the topic of regulation of virtual worlds. My notes from the presentations are below. See Mark Bell’s recap too.

Greg Lastowka, Rules of Play

Greg discussed how game rules can increase the fullness and beauty of life. Yet, legal rules may be too rational. For example, a traditional economic analysis would encourage the development of markets for virtual property, regardless of any EULA restrictions, because these markets would facilitate Pareto exchanges (i.e., both parties better off; no one worse off). However, lawyers can’t understand peoples’ need to live in beauty or how gameplay can facilitate this.

My comment to Greg: how much are rules of play exogenous to the players (i.e., imposed from the top down by the VW providers) and how much are just codifications of rules that the community of players demand on a bottoms-up basis. At Epinions, our users demanded that we vigilantly police the conduct of other players, in many cases forcing us to impose more rules or police them more vigorously than we would have done if the choice was solely ours.

James Grimmelmann

James talked about the metaphysics of virtual objects/experiences. The VW provider has the power to determine people’s perceptions within the world. For example, if the VW provider deletes an object, everyone agrees that the object ceases to exist in the world.

Also, the VW software is proprietary, so there’s no check on a provider’s autocracy. Even if the software creation was open sourced, it still wouldn’t solve the normative determination of what’s fair to do to players.

James thinks that virtual worlds need healthy virtual governance–specifically, that there should be a public sphere within virtual worlds as a way for players to discuss the providers’ autocratic decisions.

I made two comments: (1) is there anything unusual about the metaphysics in virtual worlds? It seems to me that we have many shared hallucinations in realspace (like when our government lies to us, and we accept the lie rather than listen to our inner skepticism). (2) If the risk of provider exercise of arbitrary autocratic power is a problem, couldn’t providers outsource an auditing function to third parties? For example, accounting firms audit the financial statements of companies.

Tyler asked what’s so good about liberty and fairness? I think he was driving at the fact that VWs could in fact involve benevolent dictators, and this could lead to better outcomes than we could accomplish in the real world.

There was a good Q for James that if virtual property is real, does that mean that cybertorts committed in the virtual world equally “real”? For example, is hate speech in a virtual world just as tangible as virtual property?

Tyler Ochoa, Who Owns Avatars?

Tyler acknowledged that the first response to this topic is that avatar ownership should be determined by the EULA. He made his arguments why the default ownership of avatars matters:

1) EULAs may not be binding (see, e.g., the Bragg decision, which put Second Life’s EULA in serious jeopardy)

2) some things can’t be assigned by contract, such as the 17 USC 203 termination of transfer right. (So, in 35 years, someone might come back and demand the copyrights to their avatar!)

3) the default copyright rules may determine the applicability/enforceability of contract rules

He noted the numerous aspects of an avatar that may be protected, including the avatar’s appearance, capabilities, behavior and communications.

He thinks the more that a provider gives choices to consumers to configure avatars, the more that the avatar looks like the creation of a user. As a result, he advocated that avatars should be thought of as a contribution to a collective work (although, depending on the facts, they could be a derivative work, a compilation or a joint work). He explained why this solved some of the problems about avatar ownership.

I asked Tyler whether the more appropriate model would be for providers to treat avatars as a specially commissioned work for hire as part of an audiovisual work. This would require providers to characterize the avatar as a work for hire in their EULA, but this seems like a complete solution for providers (maybe not for users!).

In his talk, Tyler asked about the appropriate remedies if a hacker deleted someone’s avatar. This seems like a problem outside of copyright law, but tangible property law could in theory apply. See, e.g., Kremen v. Cohen.

Tyler also asked what remedies users would have if their avatars were included in a derivative work (like a movie based on the VW). Again, copyright may or may not provide an adequate remedy, but it made me wonder if users have a publicity right in their avatar. I haven’t researched it, but I assume that the ROP can cover pseudonyms, nicknames, etc. If so, it seems like a derivative work may need permission from avatar alter egos irrespective of the copyright disposition.

Eric Goldman, Investment Decisions on a Shaky Virtual Foundation

Here are my notes from my talk. I am thinking about writing this up into a short essay, so I would gratefully welcome any comments.


Investment Decisions on a Shaky Virtual Foundation

There has been lots of discussion in literature about who owns virtual property. However, I’m more interested in *how* virtual property comes into existence in the first place because it gets created in a seemingly poor environment for investment decisions.

Obviously, some virtual property is generated as part of the ordinary course of gameplay. We generally don’t need to worry about the incentives to create this property; the gameplay provides the needed incentive. And I believe we don’t need to protect the investment “expectations” for this property—because the gameplay provides the motivation, there are no *investment* expectations to protect. (There may still be unhappy users who feel screwed by gameplay or providers’ monkeying with the gameplay, but this seems wholly internal to the game itself).

In other cases, virtual property will be protected by default IP laws (such as copyrightable works created within the context of Second Life). These investment decisions are no different than the creation of any other IP.

Despite these two motivations, there is still plenty of other investments made on spec or with the hope of a return, and these investments can be wiped away in a moment. There can be in-world reasons, like inflation, exploits or in-game theft. More importantly, the VW’s user agreement may give the provider unlimited ability to moot the agreement, such as by kicking the user off the site or stripping the user of assets.

The most obvious example is Second Life, where there is no gameplay per se but still plenty of real-world investment capital being invested. Second Life’s EULA makes it clear that all of this investment could be wiped away at its discretion. From its user agreement:

Sec. 1.4: You agree that Linden Lab has the absolute right to manage, regulate, control, modify and/or eliminate such Currency as it sees fit in its sole discretion, in any general or specific case, and that Linden Lab will have no liability to you based on its exercise of such right.

Sec. 2.6: Linden Lab has the right at any time for any reason or no reason to suspend or terminate your Account, terminate this Agreement, and/or refuse any and all current or future use of the Service without notice or liability to you. In the event that Linden Lab suspends or terminates your Account or this Agreement, you understand and agree that you shall receive no refund or exchange for any unused time on a subscription, any license or subscription fees, any content or data associated with your Account, or for anything else.

Sec. 5.3: When using the Service, you may accumulate Content, Currency, objects, items, scripts, equipment, or other value or status indicators that reside as data on Linden Lab’s servers. THESE DATA, AND ANY OTHER DATA, ACCOUNT HISTORY AND ACCOUNT NAMES RESIDING ON LINDEN LAB’S SERVERS, MAY BE DELETED, ALTERED, MOVED OR TRANSFERRED AT ANY TIME FOR ANY REASON IN LINDEN LAB’S SOLE DISCRETION.

So investment decisions in Second Life are made on the foundation that Second Life can moot those investments at any time for any reason. This should substantially shorten the time horizon for investment return, or at least increase the discount rate of future cash flows substantially. Yet, users still make substantial/sizable investments in Second Life and other VWs with similar policies. Why?

Hypothesis #1: Users are making irrational investment decisions because (1) they don’t know the rules governing their investments, and/or (2) they apply too low a discount rate

Evidence: <1% of users read user agreements; consumers may make mistaken inferences from the user agreements (i.e., majority of users think that the existence of a privacy policy automatically means their data must be protected-see Annenberg studies and Possible policy implications: (1) improve user education, (2) match legal terms to reflect consumer expectations, (3) caveat emptor Hypothesis #2: People are making rational investment decisions because (1) they are applying appropriate discount rate and expect short-term payoffs, or (2) they are trusting appropriate exercise of provider discretion based on market forces/brand/reputation

Possible policy implications: do nothing—market is working fine. But what if people are underinvesting due to investment uncertainty? (1) Market gives providers incentives to provide greater certainty if profitable, or (2) regulatory intervention is necessary to stabilize markets.

This got me thinking about alternative situations where people make investment decisions predicated on contracts that may be terminated for convenience. In general, US law tolerates this construct and does not establish limits on, in fact, exercising contractual rights of termination for convenience. See, e.g., United Airlines Inc. v. Good Taste (Alaska Sup. Ct. 1999). Catering company gets 3 year contract to cater United Airlines flights from Alaska, but 90 day termination for convenience clause. To perform the contract, the catering company invests $1M that (apparently) was designed to be amortized over the 3 year term. Instead, United terminates for convenience after 1 year. Catering company is unable to avoid this termination and, presumably, loses some of its investment. Indeed, in most cases involving termination for convenience, parties make some investments to perform, and contract law normally stays on the sidelines.

But in the case of franchises and distributor protection laws, we restrict a vendor’s ability to terminate for convenience even if both parties agree to a termination for convenience clause (the provider must terminate for cause or pay damages). Analogies to VW investors:

• both require upfront investment predicated on long-term support from the vendor

• vendors have substantially more leverage over contracting party—in VW context, presented on take-it-or-leave-it basis.

But noticeable differences:

• franchisors/vendors get long-lasting benefit from work of franchisees/distributors—get marketing investments/building of customer base. No direct equivalent in VWs

• franchises are heavily regulated investment decisions

More generally, does it still make sense to restrict contract freedom among franchise/distributor contracts? Or is this just an archaic paternalism?

In the case of VWs, no reason to restrict contract freedom without evidence of a problem.

• no evidence of market failure. Investments still growing rapidly

• We can rely on existing consumer protection laws (such as false advertising) provide substantial protection for any VW provider deception