Position: The Alignment Community Is Unintentionally Building a Censor's Toolkit

The Argument

"Whoever defines 'intentions and values' in AI alignment fundamentally determines whether the resulting system serves safety — or oppression."

"We do not argue for halting alignment research — its necessity is clear. Rather, we contend that given current societal, economic, and political developments worldwide, the alignment community must actively consider how our methods can be weaponized, not just perfected."

"Two entities are currently positioned to dictate AI behavior at scale: state actors and foundation model providers."

01

Alignment Methods Are Dual-Use

Pre-training filtering, RLHF, and inference-time classifiers are purpose-agnostic tools. Whoever controls them determines the values and knowledge implemented in AI systems. There is nothing in the technical methods themselves that guarantees benevolent use — and this fact has gone largely undiscussed in our community.

02

Dual-Use Is Already Happening

We systematically discuss the dual-use potential of technical alignment methods and map documented misuse across all three alignment layers — pre-training, post-training, and inference-time. To our knowledge, this framework has not been built before.

03

The Window To Act Starts Now

Growing reliance on AI for information, LLM market concentration creating power asymmetries, and global democratic backsliding to 1985 levels converge to make this moment both uniquely dangerous and uniquely important to address as a community.

Chinese models like DeepSeek and Baidu's Ernie Bot are made to refuse discussions of politically sensitive topics. As a first step in the pipeline, Chinese AI companies filter out “problematic” information and build a dataset of keywords that violate socialist values. The government additionally builds its own training datasets ("mainstream values corpus") and has blocked access to Hugging Face in 2023. China's cyberspace regulator CAC mandates model providers to prepare refusal datasets of 5,000–10,000 prompts, roughly half targeting political ideology and criticism of the communits party.

Elon Musk publicly announced plans to "fix" outputs from Grok with which he disagreed. The resulting shifts were traced to system-prompt edits — fast, cheap, and reversible. The model was steered to promote his political agenda, including claims about an alleged white genocide in South Africa. Some updates additionally triggered antisemitic responses and Holocaust denial, demonstrating the catastrophic side-effects possible when inference-time alignment is applied without oversight. This case illustrates the most accessible misuse pathway: no retraining required.

Studies show that Western LLMs apply self-censorship when prompted in Simplified Chinese — without having been designed to. The mechanism is pre-training data contamination: years of Chinese government filtering of the public internet have structurally shaped the available Simplified Chinese corpus. This means that misuse of alignment methods at one point in the ecosystem can propagate invisibly to models developed elsewhere, without any deliberate act by those developers.

Verifiable Alignment

We call for public oversight and control of alignment mechanisms. Standardized, independent benchmarks for information suppression and political bias — covering political contexts worldwide, accounting for authoritarian tendencies next to the right-left continuum, and remaining dynamic to reflect citizens' evolving realities. Users should be able to verify what a model has been made to suppress and which values it was aligned with.

Model Pluralism

No single model can achieve full neutrality. Genuine market competition prevents the dangerous concentration of informational power. Just as journalism requires diverse voices, so does AI. Monopolies over information sources are dangerous regardless of who holds them.

Awareness

Just as we advocate for user literacy about AI risks, we alignment researchers must also demonstrate our own literacy: We must reflect on and communicate the potential risks of our work more genuinely, including in our publications. Taking the impact statements seriously is a first step into that direction.

Oral Presentation

Oral Presentation 4C

Poster Session

Poster Session 5