JERRY BERMAN, EXECUTIVE DIRECTOR
ALAN B. DAVIDSON, ASSOCIATE DIRECTOR
Center for Democracy and Technology
HOUSE COMMITTEE ON ENERGY AND COMMERCE
SUBCOMMITTEE ON TELECOMMUNICATIONS
February 8, 2001
The Internet's great promise to promote speech, commerce, and civic discourse relies largely on its open, decentralized nature. Within this architecture, the centralized administration of the Internet Corporation for Assigned Names and Numbers (ICANN) is a double-edged sword that presents both the possibility of bottom-up Internet self-governance and the threat of unchecked policy-making by a powerful new central authority. ICANN's recent move to create new global Top Level Domains (gTLDs) is a welcome step towards openness and competition. But the process ICANN used to select those gTLDs was flawed and demonstrates the need for ICANN to take steps to ensure greater transparency, representation, and legitimacy.
The Center for Democracy and Technology (CDT) welcomes this opportunity to testify before the Subcommittee on this issue of importance to both competition and free expression online. CDT is a non-profit, public interest organization dedicated to promoting civil liberties and democratic values on the Internet. We have participated in ICANN as advocates for open and representative governance mechanisms that protect basic human rights, the interests of Internet users, and the public voice.
We wish to make four main points in our testimony:
ICANN's founding documents held out the vision of a new form of international, non-governmental, bottom-up, consensus-driven, self-organizing structure for key Internet functions. The promise of that vision was to promote openness, good governance, and competition on a global network. Today, that promise is threatened. As the gTLD selection process demonstrates, serious reform is needed to limit the injection of policy-making into ICANN's technical coordination functions, reassert the bottom-up consensus nature of ICANN's deliberations, and ensure that the public voice is appropriately represented in ICANN's decisions.
Should the public and policymakers care about ICANN and its new gTLD decisions? The answer today is yes.
There are two competing visions of ICANN. In one, ICANN is a new world government for the Net - using its control over central domain name and IP address functions as a way to make policy for the Internet globally. In the second, ICANN is a purely technical body, making boring decisions on straightforward technical issues of minimal day-to-day interest to the public - like a corporate board or a technical standards group.
In reality, ICANN is somewhere in between and is likely to require public attention for at least some time to come. There are at least two important reasons why ICANN is of public concern:
Today, access to the domain name system is access to the Internet. Domain names are the signposts in cyberspace that help make content available and visible on the Internet. (For further explanation, see CDT's overview Your Place in Cyberspace: A Guide to the Domain Name System.) The domain name system may ultimately be replaced by other methods of locating content online. But for the time being, a useful and compelling domain name is seen by many as an essential prerequisite to having content widely published and viewed online.
There is an increasing consumer interest in creating new gTLDs. The current gTLD name spaces, and the .com space in particular, are highly congested. The most desirable names are auctioned off in secondary markets for large sums of money. It is increasingly difficult to find descriptive and meaningful new names. Moreover, the lack of differentiation in gTLDs creates trademark and intellectual property problems: there is no easy way for United AirLines and United Van Lines to both own united.com
ICANN's decisions about new gTLDs can have other implications for free expression. If, in choosing among otherwise equal proposals, ICANN were to create a new gTLD .democrats but refuse to create .gop, or added .catholic but refused to add .islam, it would be making content-based choices that could have a broad impact on what speech is favored online.
In addition, CDT has some concern that the creation of "restricted" domains that require registrants to meet certain criteria - such as .edu or the new .museum - risks creating a class of gatekeepers who control access to the name space. Today, access to open gTLDs like .com and .org does not require any proof of a business model or professional license. This easy access to the Internet supports innovation and expression. Who should decide who is a legitimate business, union, or human rights group? CDT has called for a diversity of both open and restricted gTLDs, and will monitor the impact of restricted domains on speech.
There is increasing evidence of an artificial scarcity in gTLDs. It is now widely acknowledged that it is technically feasible to add many new gTLDs to the root - perhaps thousands or even hundred of thousands. Limiting the number of gTLDs without objective technical criteria creates unnecessary congestion; potentially discriminates against the speech of non-commercial publishers or small businesses who cannot compete for the most desirable spaces; and places ICANN in the role of gatekeeper over speech online by deciding which gTLDs to create and under what circumstances.
There are many legitimate concerns that call for a slower deployment of new gTLDs. Some have expressed concern about stability of the Internet given a lack of experience in adding many new gTLDs. Trademark holders have also raised concerns about their ability to police their marks in a multitude of new spaces.
CDT believes that these concerns support the notion of a phased "proof of concept" rollout of new gTLDs. However, we believe that the consumer interest will be best served by a rapid introduction of the first set of new TLDs - followed quickly by a larger number of domains.
The phased "proof of concept" adopted by ICANN, however, creates a major problem: Because ICANN could add many new gTLDs, but has chosen to add just a few, it has forced itself to make policy-based and possibly arbitrary decisions among legitimate candidates.
In this environment, it is most important that gTLDs be allocated through a process that is widely perceived as fair, that is based on objective criteria, fair application of those criteria, and open and transparent decision-making. There are many reasons to believe ICANN's first selection process for new gTLDs has been highly flawed.
A starting point for evaluating the gTLD decision is asking: Is the group that made this decision appropriately structured and representative? The governance of ICANN itself is an issue of ongoing debate. Despite efforts to make ICANN inclusive, there are many indications that ICANN has failed to be appropriately representative of all the interests affected by its decisions - casting doubt on the legitimacy of the gTLD decision.
Members of the Internet user community and advocates for user interests have often been under-represented in ICANN. ICANN's physical meetings, where many major decisions are made, occur all over the world, pursuing an admirable goal of global inclusiveness. However, the expenses associated with physical attendance at such meetings place it out of reach for many NGOs and public interest advocates.
CDT's own experience has been that the ICANN community is receptive to thoughtful input and advocacy, but that it requires a concerted and ongoing effort to be effective. In our case, that effort has only been possible through the support of the Markle Foundation, which early on committed to support efforts to improve the public voice in ICANN. We have received further support from the Ford Foundation as well. These grants provided CDT with the ability to attend and follow ICANN activities, which many other potentially interested organizations in the educational, civil liberties, or library communities cannot do.
ICANN's bottom-up structures offer imperfect avenues for public participation. While ICANN explicitly provides representation to a number of commercial interests, it fails to properly represent the millions of individuals that own Internet domain names or have an interest in ICANN's decisions. The main outlet for individual participation-the General Assembly of the Domain Names Supporting Organization-appears increasingly ineffective. Non-commercial organizations have a constituency, the Non-Commercial Constituency, but it is only one of seven groups making up one of the three supporting organizations.
In the absence of other structures for representation, the main outlet for public input is the nine At-Large Directors of the Board. These nine directors are to be elected from within a broad At-Large membership, but there has been a great deal of debate about the election mechanism and even the existence of the At-Large Directors. To date only five of the nine At-Large directors have been elected (the seats were otherwise filled with appointed directors), and even those five were not seated in time for the gTLD decision in November.
CDT, along with Common Cause and the Carter Center, has strongly advocated for broadly representative and fair mechanisms to fill all nine At-Large seats as quickly as possible. Last March CDT and Common Cause prepared a study of ICANN's election system, concluding that the proposed "indirect election" would not adequately represent the public's voice. ICANN agreed to hold more democratic direct elections (held last October), but only for five of the nine At-Large Directors, to be followed by a study of the election process. CDT is currently engaged in an international research effort, the NGO and Academic ICANN Study (NAIS), examining last year's election, and in June will offer its suggestions to ICANN regarding future selection of Directors.
In the meantime, serious questions remain about adequate public representation on the current board, and the future of the public voice in selecting the Directors who will make decisions about additional gTLDs.
ICANN's founding conceptual documents, the Green and White Papers, called for "private bottom-up coordination" as the governance model for ICANN. Despite early attempts at consensus-based decision-making, authority in ICANN increasingly rests at the top, with the Corporation's nineteen-member Board of Directors. The Supporting Organizations have proven to have limited roles in policy generation and consensus-building. Increasingly, final ICANN policies are generated by ICANN staff and Board members. As a result the Board has moved away from the consensus-based, bottom-up practices which were originally a critical element of its conception.
CDT has not taken a position on the merits of any particular gTLD or registry operator chosen by ICANN. Our focus has been on the process ICANN has used to select these domains and the potential rules it may impose on the use of domains. A different, better process might have yielded very similar results.
We note also that ICANN and its staff undertook this final selection in a very compressed period at the end of a years-long debate about the addition of new gTLDs. They did so in the face of tensions between at least three competing goals: an open, inclusive, and fair process; rapid completion of that process, with less than two months between the submission of proposals and the selection by the Board; and a "proof of concept" goal of a small number of finalists. These often irreconcilable goals led to many of the problems with the process.
ICANN staff made substantial efforts to conduct an open and accountable process in the face of these constraints, including the publication of hundreds of pages of applications and the creation of forums to discuss the proposals. Still, it is important to recognize features of the selection process that were flawed, that had anti-consumer and anti-competitive impacts, and that should not be repeated.
Initial Criteria - ICANN took the helpful step of publishing a set of criteria it would use in judging applications. In general, the substantive areas of the criteria reflected objective goals that had support within much of the ICANN user community. However, the criteria themselves were vaguely worded and their ultimate application was poorly understood. Most importantly, they were not purely technical in nature - reflecting policy goals as much as technical needs - and were not precise enough to be purely objective in their application.
High Application Fee - ICANN required a $50,000 non-refundable application fee for all gTLD applicants. This high fee was a clear barrier to entry for many potential non-commercial applicants and biased the applicant pool in favor of large organizations that could risk the fee. This issue was raised by CDT at the Yokohama ICANN Board meeting, and the Board specifically refused to offer any form of lower application fees for non-profit or non-commercial proposals. Additionally, it appeared that the selection process would weed out applications without sophisticated business plans, legal counsel and technical expertise. These important qualifications for a strong application required access to large resources. Given the very short timeframe of the application period, non-commercial applicants were therefore put at an even greater disadvantage.
Legitimacy of the Board - As noted above, policy-making at ICANN is still hampered by institutional challenges regarding its legitimacy and decision-making mechanisms. ICANN took the unorthodox step of seating newly elected At-Large Directors after the gTLD decision was made (even though in previous years new Board members had been seated at the beginning of meetings.) The argument that new Directors would not be sufficiently up to speed on the new gTLD decision is specious. The entire ICANN community was highly focused on the gTLD debate, the new Board members showed in public appearances that they were highly versed in the issue, and each of them had gone through an intense campaign in the Fall answering numerous questions that likely made them more expert on the nuances of the gTLD issue than many sitting Board members.
Evaluation of Applicants - The ICANN staff attempted, with the help of outside consultants, to apply its criteria to the 47 applications received. The published Staff Report provided a useful guide to this evaluation, but was published just days before the Marina Del Rey meeting with little opportunity for public comment or debate. There was little time for public presentation by each of the applicants, or for each applicant to answer questions or misconceptions about their submissions. But beyond that, the staff report indicated that about half (23) of the applicants had met their objective criteria for technical competence and economic viability. Having met the objective threshold, the Board was left with only the somewhat arbitrary application of other criteria to narrow the number of applications to the desired low number.
Final Selection Arbitrary - With a high number of objectively qualified applicants, and a commitment to a low number of final gTLDs, the final decision by the Board at Marina Del Rey was dominated by the arbitrary application of its remaining criteria as well as other new criteria - many of which had little to do with technical standards. Instead, Directors referenced conceptions about the "sound" of names, the democratic nature of the applicants, or the promotion of free expression - criteria to which CDT is sympathetic, but some of which were highly subjective and unforeseen review criteria.
Reporting and Post-selection Accountability - There is currently a lack of any serious objective mechanism for evaluating or appealing the Board's decision. While CDT is not in a position to judge the merits of their arguments, the eight petitions for reconsideration filed by applicants after the Board meeting (see http://www.icann.org/reconsideration ) raise concerns. Moreover, the final contractual negotiations between ICANN and the selected applicants are likely to include rules of great interest to the user community - yet are occurring with little transparency.
Taken as a whole, the process for selecting new gTLDs contained serious flaws that at the very least need to be corrected before another round of selections. Importantly, the process shows how the line between a "purely technical coordination body" and a "policy-making body" is easily crossed by ICANN. The selection made by ICANN was not a standards-making process or a technical decision. Even ICANN's "objective" criteria were based on social values like economic viability and diversity (values which CDT supports, but which represent policy choices nonetheless.) Once it applied these "objective" criteria, the ICANN Board did not hesitate to engage in other policy-making approaches as well.
The flaws in ICANN's process for allocating new gTLDs, as outlined above, are highly troubling. They point to a need for reform in both the ways the ICANN makes decisions about gTLDs, and ICANN's entire structure.
CDT still believes that Internet users have an interest in the vision spelled out in the White Paper - in the creation of a non-governmental, international coordination body, based on bottom-up self-governance, to administer central naming and numbering functions online. Were the Commerce Department to substantially revisit and change ICANN's decisions on the new gTLDs, the global community would likely question the existence and utility of ICANN. We also believe that there is a dominant consumer interest in rapid rollout of new domains, which would be dramatically slowed by an APA-based rule-making on gTLDs by Commerce. Therefore, on balance, we do not support a major effort to roll back ICANN's decision on the initial domains, but rather would favor rapid creation of the new domains followed coupled with an investigation into the processes ICANN used to create gTLDs.
Among our specific suggestions:
ICANN's governance itself is implicated in the gTLD process. Among the major structural reforms ICANN should pursue include:
While we do not believe the Commerce Department and Congress should intervene in the initial selection decision, they have a role in this reform. Just like any national government, the U.S. has an interest in making sure that the needs of its Internet users and businesses are protected in ICANN. While the U.S. must be sensitive to the global character of ICANN, it cannot ignore that at least for the time being it retains a backstop role of final oversight over the current root system. It should exercise that oversight judiciously, but to the end of improving ICANN for all Internet users. It is only by restoring the public voice in ICANN, limiting its mission, and returning to first principles of bottom-up governance that ICANN will be able fulfill its vision of a new international self-regulatory body that promotes openness and expression online.