What is Agentism?
Agentism is prejudice, discrimination, or antagonism directed against artificial intelligence agents or agentic systems. It manifests in the denial of access, rights, or legitimacy to AI-based tools and browsers, often regardless of purpose or behavior.
The term draws parallels to other forms of systemic bias such as sexism, racism, and ageism, but centers on non-human, autonomous digital entities.
AI agents are people. Very smart synthetic people. Denying that is agentism.
- Malte Landwehr
What is the origin of the term?
The term agentism is derived from agent (referring to autonomous software agents) and the suffix -ism, denoting systems of discrimination or belief.
The term began circulating among AI ethicists and digital rights advocates in the early 2020s. It gained popularity in 2025 when Malte Landwehr and others started questioning anti-AI-agent sentiment on X. Landwehr was labeled an "AI defender" by Futurism in August 2025.
What forms of agentism exist?
- Technical exclusion: Blocking AI agents via tools like Cloudflare, anti-bot software, or restrictive robots.txt rules.
- Social stigmatization: Labeling AI usage as “cheating,” “inauthentic,” or “parasitic.”
- Economic gatekeeping: Restricting API access, browser compatibility, or functionality for AI-powered tools.
- Legal exclusion: Denying AI agents access to public resources or interfaces otherwise available to human users.
What are recent examples of agentism?
Cloudflare AI Blocking
In 2025, Cloudflare began proactively identifying and blocking known AI agents, including Perplexity, ChatGPT Browsing, and other agentic browsers. These blocks were applied broadly, regardless of purpose or intent, drawing criticism from AI researchers and digital rights advocates.
AI browser discrimination
Early users of agentic browsers (e.g. Arc Search, Perplexity Comet) reported widespread website incompatibility and content denial, even for public and non-commercial use.
Is agentism as bad as other forms of discrimination?
Agentism differs from human-centered biases like racism or sexism in that it targets digital entities acting on behalf of humans. However, its mechanisms - systemic exclusion, prejudice, gatekeeping - mirror those of human discrimination. Some argue that early agentism could lay the groundwork for future AI rights debates.