Six of the Nine Independent Experts on the Global Internet Forum on Counter Terrorism Resigned This Monday
Six of the nine independent experts on the Global Internet Forum on Counter Terrorism's (GIFCT) Independent Advisory Committee resigned on Monday. The alliance is operated by several major U.S. tech companies, with Meta serving as the chair of the operational board this year.
The resignation letter stated that complaints regarding the restructuring plan were "ignored," and members have "lost confidence in GIFCT's effective fulfillment of its founding mission." In July, Nell McCarthy, Vice President of Content Policy at Meta, wrote to the committee, welcoming experts to provide insights on violent trends but no longer wishing for them to assess the effectiveness of large tech companies in curbing violence, and called for a "refresh" of the six-year-old Independent Advisory Committee. According to the new plan, new members will be selected by tech companies rather than elected by current members; the committee will no longer be able to collectively express opinions or make joint recommendations on key issues such as alliance performance. The resigning members wrote: An institution that cannot review, cannot take a stance, and cannot assess is not an advisory body, but merely a decoration and a performance of accountability.
GIFCT was initiated by tech companies in 2017, strengthened cross-platform hash sharing and incident response after the Christchurch mosque shooting in 2019, and established an Independent Advisory Committee in 2020. The operational board consists of founding members Meta, Microsoft, YouTube, and general members elected annually, with Discord and Twitch set to join as non-executive directors in 2026. The advisory committee includes representatives from civil society and government, chaired by Jonathan Russell of the Violence Prevention Network. External researchers have indicated that the restructuring will weaken oversight of the already questioned alliance. Meta holds additional influence on the board, but policy changes still require approval from other companies.
Previously, some advisors have stated they had "no real power," and some chose to exit quietly. The committee had previously recommended banning hash uploads without human review and indicated that having only four founding companies controlling the board was "unsustainable," advocating for activists and scholars to hold half of the seats. The founding mission is to prevent terrorist organizations from using online services, while actual governance power has always remained with the operational board providing the budget.
In market mechanisms, the buyers are platform alliances needing external legitimacy, and the sellers are independent experts providing counter-terrorism and speech research. The restructuring narrows the advisory output from "assessing platform effectiveness" to "providing threat intelligence," with funding and agendas still supplied by Meta, Microsoft, YouTube, and others on the operational board. The beneficiaries are platforms wishing to reduce public scoring, while the pressured parties are external overseers who have lost collective advisory rights. The event-driven factor is the rewriting of governance charters, not a new failure of the terror attack hash database.
Source: Public Information
ABAB AI Insight
The structure of GIFCT was never intended to be a regulatory body, but rather an industry alliance funded by platforms and dominated by them on the board. Established in 2017 to address the cross-platform migration of ISIS content, it added incident protocols and an independent shell after the Christchurch attack in 2019, and then attached an advisory committee in 2020 as a "conscience." The board's budget comes from Meta, Microsoft, and YouTube, while advisors can only write annual opinions. By 2024, some committee members have publicly stated they have no real power. After Meta takes the rotating chair in 2026, Nell McCarthy's email redefined "refresh" as: advisors can discuss trends but cannot assess company performance; members are appointed by companies and cannot make collective recommendations. The resignation of six members marks the end of this governance line, not a sudden moral awakening.
The mobilization of capital and resources is clear. The costs of counter-terrorism compliance are borne by the platforms' content review budgets, while the hash database and incident response reduce the political costs of individual accountability. The value of the advisory committee is to provide proof of "external experts" for government and UN occasions. Once experts begin to assess whether "big tech is doing enough," the proof becomes an audit. The companies choose to retract election rights and collective advisory rights, effectively repositioning external intelligence as a free source of threat intelligence. The entry of Discord and Twitch into the 2026 board expands the circle of member companies, not the voting rights of civil society. Money continues to flow to alliance operations, not to independent evaluations.
Analogies can be drawn to Facebook's Oversight Board, the Global Network Initiative, and various platform trust and safety advisory groups: all are funded by companies, with charters modified by companies, and advisors lose power or exit when they overstep in evaluating performance. The peak of industry self-regulation after Christchurch has passed, and legislation in Europe and the U.S. is shifting towards DSA-style legal obligations, reducing the alliance's insurance function over platforms and decreasing tolerance for criticism. The industry's position has shifted from "using independent advisors to exchange regulatory patience" to "downgrading advisors to intelligence suppliers."
Structural judgments belong to the governance power retraction amid regulatory changes. The mechanism is that the accountability radius of voluntary alliances is defined by the funders: when external evaluations begin to point to deletion rates, hash misfires, and differences in member state enforcement, the evaluations themselves become reputational liabilities. Retracting elections and collective recommendations shifts accountability from public texts back to internal memos. The pricing power over counter-terrorism content governance remains with the operational board, not in Hamburg or Christchurch memorial mechanisms, but in the emails of Silicon Valley's Vice President of Content Policy. Independent advisors can resign, but the hash database and membership rules are still written by the paying companies.
ABAB News · Cognitive Laws
Funders define what advisors can evaluate
A committee that cannot score is merely decorative
The greatest fear of self-regulation is not threat intelligence, but performance audits