What’s in an “AI Name”? The Psychology, History, and Future of Naming Artificial Intelligence

When Joseph Weizenbaum created ELIZA at MIT in 1966, he gave humanity its first conversational computer program. By naming his script after Eliza Doolittle—the fictional flower girl from Pygmalion who learns to speak proper English—Weizenbaum accidentally started a tradition that persists to this day: giving algorithms human-like names to soften the cold reality of lines of code.

Decades later, the question of what to call an artificial intelligence system has evolved from a whimsical programmer’s choice into a multi-million-dollar branding science. From the vintage utility titles of early virtual assistants to the sleek, abstract monikers of modern generative models, an AI’s name shapes how we trust, use, and perceive technology.

1. The Three Eras of AI Naming Conventions

The taxonomy of AI names has shifted dramatically alongside technological breakthroughs. Looking back at the evolution reveals how our relationship with machines has changed:

• The Anthropomorphic Era (The 1960s–2000s): ELIZA, PARRY, and later A.L.I.C.E. (Artificial Linguistic Internet Computer Entity). These names were explicitly human or acronym-driven to mask mechanical limitations. They invited users to lean into the illusion of empathy, often mimicking psychological profiles.
• The Assistant Era (The 2010s): Siri, Alexa, Cortana, and Google Assistant. Tech giants favored friendly, crisp, mostly female-coded names that felt subordinate, helpful, and easily roll off the tongue for voice commands. They were designed to evoke a digital butler waiting in the wings.
• The Oracle & Abstract Era (Present Day): ChatGPT, Claude, Gemini, Midjourney, and Copilot. Today’s naming conventions lean heavily toward minimalist functionality, elevated abstraction, or literary nods. Rather than pretending to be a human friend, modern AI branding leans into being a collaborator, tool, or co-pilot.

2. The Psychology of Humanizing Code

Why do tech companies rush to give human names to complex neural networks? It comes down to the ELIZA effect—our cognitive tendency to unconsciously attribute human emotional depth and understanding to computer behaviors.

When an AI has a clean, approachable name:

1. Friction Drops: People are far more likely to adopt and experiment with a tool called “Claude” or “Siri” than they are with an internal engineering project code-named Model-7B-v4.
2. The Trust Paradox: A human-like name builds initial rapport, but it can also backfire. If an AI possesses a warm, personal name, users are more likely to feel betrayed or frustrated when it hallucinates facts or makes basic reasoning errors.
3. Cultural Resonance: Names like Gemini (signifying duality and multi-modal capabilities) or Copilot (signifying partnership rather than replacement) communicate a system’s core philosophy before a user even types a single prompt.

3. The Mechanics of Branding a Modern AI

Behind every major AI release is a grueling naming matrix designed to survive global trademark laws, cultural translation tests, and linguistic barriers. Brand strategists look at three primary buckets:

• Descriptive: Names that tell you precisely what the system does (