ChatGPT and Google Bard are both large language models (LLMs) that are trained on a massive dataset of text and code. This means that they are both capable of generating human-quality text, translating languages, writing different kinds of creative content, and answering your questions in an informative way.

One of the reasons that ChatGPT and Google Bard produce articles that are so very similar is that they are trained on a similar dataset of text. This dataset includes news articles, books, code, and other forms of text. As a result, both models have learned the same patterns and rules for how to generate text.

Another reason for the similarity of their output is that both models are trained on a supervised learning task. This means that they are given a large number of examples of text and code, and they are asked to generate similar text and code. This training process helps the models to learn the patterns and rules that are most likely to produce human-quality text.

Finally, both ChatGPT and Google Bard are constantly being improved. As they are exposed to more data and more examples of text and code, they become better at generating human-quality text. This ongoing improvement is one of the reasons that the output of these models is becoming increasingly similar to human-written text.