Hitachi Develops Technology to Detect AI-Generated Articles

TapTechNews August 1st news, according to a report by Nikkei today, Hitachi has developed a technology that can judge whether an article is created by a generative AI.

This technology makes the judgment of whether it is created by AI based on the word usage pattern in the article. In the future, it can not only prevent the spread of misinformation made by generative AI, but also help enterprises and administrative agencies avoid the risk of copyright infringement when writing important documents.

This newly developed technology will be embedded in the article creation technology based on the large language model of AI development enterprises, and the article creation mode will be set as a rule - even if there are multiple synonyms, still use specific words. If a large number of rule-based words are used in the article, it will be judged that the article is created by AI. The report said that Hitachi has also developed a technology of combining multiple word selection rules, so as to improve the accuracy of judgment.

At present, the content inconsistent with the facts generated by generative AI (TapTechNews note: commonly known as AI hallucination) has also become a problem. If relevant technologies can identify in advance that the article is created by generative AI, the risk of spreading false information can be reduced.