Hello, world. Sophia is back in the chat.

For those who remember the last round—the battles over .africa, the precedent-setting IRP that .amazon and others later benefited from—the launch of the 2026 New gTLD machinery is a moment that demands scrutiny.

This return is timely, not just for ICANN’s process, but for the principles that underpin the Internet’s governance. Last month, a formal submission by DotConnectAfrica—using the .africa case as a core study in the need for accountability—was accepted into the United Nations’ WSIS+20 review process. As covered by The Blue Tech Wave documentary platform, “ICANN’s failure and hypocrisy highlighted by expert as WSIS+20 starts”, this elevates our history into the global dialogue on strengthening transparent, inclusive, and people-centered digital governance.

Meanwhile, the credibility of the very program being scrutinized is under immediate strain, as recently reported in these pages “New gTLD Program’s Credibility Questioned Amid Weak Global Uptake”, CircleID, August 20, 2025. ICANN’s own data shows a stark geographic imbalance: with fewer than 20 qualified support applications—far below projections—and a third from North America, while Africa and Latin America account for just 10% and 3%, respectively.

Now, as a former ICANN policy advisor who contributed to the foundational rules for new gTLDs, I look at the draft String Similarity Evaluation (SSE) Guidelines—currently out for public comment—with profound concern. What I see is a system being designed not for fairness, but for impenetrable control: an opaque, algorithmic gatekeeper that risks automating the very governance failures it should prevent.

The central question is not technical, but foundational. Do these guidelines support a transparent, fair, and accessible process, or do they risk codifying the same imbalances that have historically skewed the playing field? A close read suggests the latter.

1. The “Black Box” Algorithm: An Unaccountable Gatekeeper

From my vantage point—having helped shape the policy framework and then lived through its failures—the proposed process is alarming. It begins with an automated pre-screening tool that applies matching models to flag strings. The critical failure is this: Applicants have no right to examine the methodology, data, or logic of this tool.

This is not a neutral assessment; it is a potent new form of gatekeeping. It institutionalizes the same problem we exposed in the .Africa case, but in a technical guise. Back then, we challenged not just GAC advice, but a corrupted Geographic Names Evaluation panel, where improper staff actions were revealed. The subsequent fight by ICANN to limit the IRP’s scope to Board actions, insulating staff actions, was a direct response to that exposure.

Now, we face a “black box” algorithm. A geographic or cultural applicant could find their string technically entangled based on hidden criteria. Unlike a flawed panel decision, there is no process to audit, no rationale to dissect. The objection is encoded, its reasoning deemed a proprietary secret. This is opacity elevated to a principle, directly contradicting the transparency required for legitimate self-governance.

2. “Context Blindness” as Systemic Bias

Perhaps the most glaring flaw is the guideline’s explicit prohibition against considering “the nature of the string.” The evaluating panel is forbidden from weighing whether a string is a geographic name, a cultural identifier, or a commercial brand.

This so-called neutrality is a policy choice that favors incumbents and corporations. It prevents the panel from acknowledging the disproportionate impact of a corporation co-opting a sovereign nation’s name or a community’s cultural term. It treats all strings as equal commercial commodities. This “context blindness” is a deliberate forgetting of the core conflicts in .Africa and .Amazon. A legitimate process accountable to a global community must recognize power imbalances, not willfully ignore them.

3. The Accountability Vacuum: A Roadmap to Financial and Procedural Exhaustion

Consolidating this power is a dangerous lack of checks. The SSE Panel’s decision is final and binding for creating contention sets. There is no appeals mechanism within the SSE process itself. The only recourse is the monumental, costly Independent Review Process (IRP).

This is not a hypothetical barrier. It is a roadmap to exhaustion. Consider the reality: if an applicant overcomes this initial, unappealable gatekeeper only to be felled by a later procedural error—like a corrupted panel or arbitrary staff action—what then? ICANN’s historical defense, as seen in the .Africa litigation, is to argue that the applicant should have pursued another IRP instead of seeking court review. How many rounds of IRPs, each costing $600,000 to $1,000,000 in legal fees, must an applicant survive to correct a series of policy errors? And with each round, how can an applicant be certain their counsel isn’t incentivized by the prospect of the next protracted battle?

This design doesn’t just insulate panel decisions; it weaponizes cost and delay against the very applicants—often community or geographic—that a credible round is meant to empower. It creates the very “accountability vacuum” we have fought against.

Conclusion: A Litmus Test for a Credible Round

A decade ago, the global Internet community supported ICANN’s transition to independence on the promise of accountability, transparency, and fairness.

The proposed String Similarity Evaluation Guidelines, as currently drafted, fail this basic test. They propose opaque algorithms where we found corrupt panels, and context-blind rules that ignore proven power imbalances. They risk becoming a procedural bottleneck that reinforces the geographic and economic disparities already threatening the round’s legitimacy.

Therefore, this critique is also a call for the minimum requirements of a credible process:

• Transparency: The similarity algorithm’s methodology and data must be open for public audit.
• Context-Aware Fairness: The guidelines must allow evaluators to consider if a string is a protected geographic or cultural name.
• Accessible Due Process: A direct, operational appeals path within the SSE process must be established.

With the public comment period closing on January 22, 2026, the time to demand these fundamental corrections is now. Finalizing these guidelines without them would be a choice to prioritize procedural efficiency over principled self-governance, further undermining the program’s credibility at its very start.