How to beginner · 3 min read

How to detect AI-generated content

Quick answer

Detect AI-generated content by analyzing text with specialized models like gpt-4o or claude-3-5-sonnet-20241022 using prompt engineering or dedicated detection APIs. You can also use statistical or embedding-based methods to identify AI writing patterns.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0

Setup

Install the openai Python package and set your API key as an environment variable.

Run pip install openai
Set environment variable OPENAI_API_KEY with your OpenAI API key

bash

pip install openai

Step by step

Use the OpenAI gpt-4o model to classify if a text is AI-generated by prompting it directly. The example below sends a prompt asking the model to detect AI-generated content and returns the model's assessment.

python

import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

text_to_check = "This is a sample text to test if it was generated by AI or a human."

prompt = f"Determine if the following text is AI-generated or human-written. Respond with 'AI-generated' or 'Human-written' only.\n\nText:\n{text_to_check}"

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": prompt}]
)

result = response.choices[0].message.content
print(f"Detection result: {result}")

output

Detection result: AI-generated

Common variations

You can use the Anthropic claude-3-5-sonnet-20241022 model similarly by sending a system prompt and user message to detect AI-generated content. Alternatively, use async calls or stream responses for large-scale detection.

python

import os
import anthropic

client = anthropic.Anthropic(api_key=os.environ["ANTHROPIC_API_KEY"])

text_to_check = "This is a sample text to test if it was generated by AI or a human."

system_prompt = "You are an AI content detector. Respond with 'AI-generated' or 'Human-written' only."

messages = [{"role": "user", "content": f"Detect if this text is AI-generated:\n{text_to_check}"}]

response = client.messages.create(
    model="claude-3-5-sonnet-20241022",
    max_tokens=50,
    system=system_prompt,
    messages=messages
)

print(f"Detection result: {response.content[0].text}")

output

Detection result: AI-generated

Troubleshooting

If the detection result is ambiguous or inconsistent, try refining the prompt for clarity or increasing max_tokens. For API errors, verify your API key and network connectivity. Detection accuracy varies by text length and style.

✅

Key Takeaways

Use gpt-4o or claude-3-5-sonnet-20241022 models with clear prompts to detect AI-generated content.
Set environment variables for API keys and install official SDKs for reliable integration.
Prompt engineering is critical for accurate classification; keep prompts explicit and concise.
Detection accuracy improves with longer text samples and context.
Check API keys and network if you encounter errors or unexpected responses.

Verified 2026-04 · gpt-4o, claude-3-5-sonnet-20241022

Verify ↗