Exploring the Limits of AI Copywriting

An Experiment in Creating a Crypto News Website


As an Elixir developer interested in coding and experimenting, I decided to embark on a unique project. My goal was to create a website to parse news articles from another crypto news site and use AI to copywrite texts from the parsed articles. The key idea of this experiment is to explore the potential of AI-generated content in attracting search engine traffic.

In this article, I will outline the process of creating this website, discuss the ethical concerns and challenges involved, and share my findings on the viability of using AI-generated content for search engine optimization (SEO) and monetization.

Ethical Concerns and challenges

As we dive deeper into the world of AI-generated content, several fears and ethical concerns arise. Here, I will address these fears and discuss how they pertain to my experiment:

  1. Difficulty in distinguishing AI-generated content from human-created content: As AI-generated content becomes increasingly sophisticated, it becomes harder to differentiate it from content created by human writers. This raises concerns about the potential for AI-generated content to mislead readers and diminish the value of creative work.
  2. Proliferation of AI-generated content: With the increasing accessibility and affordability of AI-driven content-generation tools, the internet might soon be flooded with such content. This could dilute the overall quality of content and make it harder for users to find accurate and reliable information.
  3. Validating content authenticity: As AI-generated content becomes more prevalent, developing methods for validating content authenticity becomes essential. Solutions like digital watermarks or blockchain-based content verification could be vital in addressing this concern.
  4. Misinformation on the internet: The internet is already filled with incorrect information, and the proliferation of AI-generated content could exacerbate this issue. Ensuring that AI-generated content is factually accurate and reliable is an ongoing challenge.


  1. I’m going to use several site donors, parse them every 15 minutes and save them into a table, where we will store parsed posts. For this purpose we will use scheduler module, that will spawn Oban jobs for each parser.
  2. There will be a separate process, that will check table with temporary posts. When it gets a new temporary post, it will use Copywriter API, rewrite content of the post and create a real post in another table. These posts will be listed on the site.
  3. Looks quite simple.


Firstly we create behaviour, to be sure that parse/0 is defined in every parser module.

Our parser module:

Oban worker:

Scheduler that will create parsing jobs in Oban:

Experiment Outcomes and Monetization potential

Upon completing my crypto news website and implementing AI-generated content, I intend to monitor its performance regarding search engine traffic, ad revenue, and affiliate marketing. Here is a list of metrics that I’ll check:

  1. Search engine traffic - how attractive is AI generated content for search engine systems.
  2. Ad revenue - is it possible to earn money through showing Ads on the site.
  3. Affiliate marketing - is it possible to earn money through affiliate products.

In conclusion, embarking on this experiment to develop a crypto news website using AI-generated content promises to be an intriguing journey in the upcoming 2-3 months. As we move forward, I will continue to monitor the site’s metrics to evaluate its performance and gain further insights into the effectiveness of AI-generated content in attracting search engine traffic and monetization. Additionally, I will be sharing a technical article detailing the process of creating an API client specifically for copywriting purposes. This may provide valuable information to developers and enthusiasts who wish to explore the world of AI-driven content creation. Stay tuned for future updates on this fascinating project!

Next article:

Implementing AI Copywriting with ChatGPT in Elixir