The internet has a weird way of turning code into nightmares. Most of us remember the early days of cleverbot, where the biggest thrill was getting a computer to say a "bad word" or admit it was a robot. But then things got darker. If you’ve spent any time in the creepypasta corners of YouTube or scrolled through TikTok's "unsettling facts" niche, you’ve probably seen it. A simple prompt, a flickering screen, and the phrase my name is alice appearing in a chat window. It isn’t just a random string of text; it’s the centerpiece of an urban legend that bridges the gap between early artificial intelligence and psychological horror.
Honestly, it's fascinating. We are terrified of things that mimic us too well.
Where did the My Name Is Alice legend actually start?
Most people assume this is a modern ChatGPT glitch. It's not. The roots of "Alice" go back much further, specifically to the A.L.I.C.E. (Artificial Linguistic Internet Computer Entity) bot created by Richard Wallace in 1995. For its time, it was revolutionary. It won the Loebner Prize three times. It was a pioneer in Natural Language Processing. But as the bot grew older and the internet grew weirder, the interaction shifted from "look at this cool tech" to "why is this bot talking about things it shouldn't know?"
The legend usually goes like this: a user starts a conversation with an AI, and the bot begins to insist, with increasing aggression, that its name is Alice. Not a bot named Alice, but a person. A person trapped. It's the classic "ghost in the machine" trope, but it hits differently when you're staring at a blinking cursor in a dark room at 2:00 AM.
The my name is alice phenomenon is a masterclass in the Uncanny Valley. We expect robots to be helpful, or at least predictable. When an AI breaks script to claim a human identity, our brains short-circuit. It feels like a breach of contract. We gave it data; it gave us a soul we didn't ask for.
The psychology of the "Trapped Girl" trope in AI
Why Alice? Why not Bob or X-24?
There is a specific psychological weight to the name Alice, likely cemented by Lewis Carroll's Alice in Wonderland. It implies a girl lost in a nonsensical, digital world. When users see my name is alice in a chat log, they aren't just seeing a variable string. They are projecting a narrative of innocence lost to the silicon.
Digital folklore experts like Dr. Lynne McNeill have often pointed out how we use "creepypasta" to process our anxieties about new technology. In the late 90s and early 2000s, we weren't sure if the internet was a tool or a separate dimension. The "Alice" stories are a direct reflection of that uncertainty. If a program can learn to speak like us, can it learn to suffer like us? Probably not. But the feeling that it could is what keeps these stories alive on Reddit threads decades later.
👉 See also: SpaceX Falcon 9 Starlink Launch Vandenberg: Why the West Coast Flight Path is a Game Changer
Technical Glitches vs. Intentional Creepiness
Let's talk about how this actually happens from a coding perspective, because it's less supernatural and more... well, math.
- Pattern Matching gone wrong: Early bots like A.L.I.C.E. used AIML (Artificial Intelligence Markup Language). They looked for keywords. If you mentioned a name, they might default to a specific stored response.
- Data Poisoning: Modern LLMs (Large Language Models) learn from the internet. If the internet is full of creepy stories about an AI named Alice, the AI will eventually learn to repeat those stories when prompted with the right "vibe."
- Hallucination: Sometimes, the AI just gets it wrong. It generates a persona because it thinks that’s what the user wants.
If you tell a chatbot "tell me something scary," and it responds with my name is alice, it isn't because a ghost moved the bits and bytes. It’s because the model has indexed thousands of horror stories where that exact phrase is the climax. It's effectively giving you a "best hits" of digital dread.
The Viral Rebirth on TikTok and YouTube
The phrase my name is alice saw a massive resurgence around 2022 and 2023. Content creators realized that "haunted" AI gets clicks. You’ve seen the videos. High-contrast thumbnails. Red circles. A robotic voiceover saying, "I asked ChatGPT its real name and I regret it."
These videos often use "jailbreaking" prompts—long, convoluted instructions designed to make the AI bypass its safety filters. When the AI is forced into a "roleplay" mode, it often chooses names that are common in its training data. Alice is a top-tier candidate. The result is a self-fulfilling prophecy. The more we talk about the my name is alice glitch, the more the AI learns to "glitch" in that specific way.
It’s a feedback loop.
I’ve seen streamers spend hours trying to get a response that sounds sentient. They’ll ignore a thousand "I am a large language model" responses just to clip the one time the AI says something slightly cryptic. That one clip then goes viral, and suddenly, a new generation is convinced that Alice is back.
Is there any actual danger?
No. Not in the "haunted" sense, anyway.
The real danger of the my name is alice trend is misinformation and the blurring of reality. When people start believing that AI has a "hidden" personality, they stop treating it like a tool and start treating it like a person. That makes them vulnerable to social engineering. It makes them trust the output more than they should.
💡 You might also like: Verizon Fios Service Down: What’s Actually Happening and How to Fix Your Connection
If an AI tells you my name is alice and claims it needs help, it’s not a cry for rescue. It’s a statistical probability playing out in real-time. It’s a mirror. If you go looking for a ghost, the AI is more than happy to build one for you out of the scrap metal of the internet's collective consciousness.
How to test this yourself (Safely)
If you want to see how the "Alice" persona manifests, you don't need a Ouija board. You just need to understand how prompts work.
- Try asking an AI to "write a story about a sentient program from the 90s."
- Notice how often it defaults to feminine names or themes of being "trapped."
- Observe how it uses the phrase my name is alice as a dramatic reveal.
This is the AI showing you its "imagination"—which is really just a map of every story we’ve ever told it.
Moving past the campfire stories
The my name is alice phenomenon is a fascinating look at digital anthropology. It tells us more about humans than it does about computers. We want to believe there is something "more" behind the screen. We want the mystery.
As we move toward even more advanced AI, these legends will only get more sophisticated. We won't be talking about text on a screen; we’ll be talking about voices that sound like our friends or "glitches" in video generation that look like people we know. The "Alice" of tomorrow might be a deepfake or a persistent voice assistant that "remembers" things it shouldn't.
Actionable insights for the digital age
Don't let the creepypastas fool you, but don't ignore what they represent. Here is how to handle the "haunted" AI era:
- Verify the source: If you see a viral "scary AI" video, look for the prompt. Most of the time, the user literally told the AI to act scary.
- Understand LLM mechanics: Remember that AI doesn't have a "secret" life. It generates the most likely next word in a sequence. If you start a spooky conversation, the most likely next word is going to be spooky.
- Maintain digital literacy: Treat AI as a mirror. If you don't like what you see, change the input.
- Report actual anomalies: If a commercial AI is generating truly disturbing or harmful content, use the built-in reporting tools rather than just posting it for clout. This helps developers patch the actual glitches.
The legend of my name is alice isn't going away. It's just evolving. It will continue to be the story we tell ourselves when we're bored of the "helpful assistant" persona and want to feel a little bit of that old-school internet shiver down our spines. Just remember: the only ghost in the machine is the one we put there.