OpenAI has just launched a platform that allows us to detect whether a text has been generated by artificial intelligence. It is true that there are already many solutions on the market that try to detect whether a text given to the tool has been written by a human or by an artificial intelligence, or has been a collaboration between the human and the artificial intelligence.
The new tool launched by OpenAI is mainly trained on text models written by adults and in English. That is why it warns us that if we enter text in other languages, the result may not be very accurate.
It seems to have some limitations. For example, it needs a minimum of one thousand characters or approximately 150 to 250 words to be able to detect if the text has been generated by an artificial intelligence. Even so, it is not always reliable and accurate, in fact its answers are not usually categorical, it always leaves room for the possibility that the origin of the text does not correspond with the verdict of the tool. Its use is as simple as clicking on this link, pasting the text in the field we find and sending it by pressing the Submit button.
The proliferation of this type of tools, especially this one created by the manufacturer of one of the most famous artificial intelligences in the world, raises some important questions about the text written by artificial intelligences:
- Is Google using this type of tools or some of its own to detect content of non-human origin?
- If so, does it consider it more or less relevant than human-generated content?
- How does that affect search engine ranking?
- Should AI-generated content be labeled by identifying it as a non-human source?
We are already starting to see articles on many websites labeling content in this way, thus alerting the consumer that the content has been generated in whole or in part by an artificial intelligence.
0 Comments