 A study of global seismic noise from 1997 to 2022 was conducted using the Donohoe-Johnston DJ index, which separated the values of wavelet coefficients into small and large. For each reference point in an auxiliary network of 50 points, a time series was calculated with a time step of one day for the median of the values at the five nearest stations. Correlations between the DJ index values at the reference points were then calculated over a moving time window of 365 days. This revealed a decrease in the average values of the DJ index and an increase in correlations indicating an increase in global seismic danger. After 2011, there was a sharp increase in the maximum distances between reference points with large correlations. Additionally, the high amplitude of the response of the DJ index to the length of the day for 2020 to 2022 could predict a strong earthquake in the second half of 2023. The study aimed to improve the mathematical apparatus for assessing the current seismic hazard based on the properties of seismic noise. This article was authored by Alexei Labushan. We are article.tv, links in the description below.