<jats:title>Abstract</jats:title><jats:p>Although evidence of recovery in Antarctic stratospheric ozone has been found, evidence of recovery in Arctic ozone is still elusive, even though 25 years have passed since the peak in ozone depleting substances. Here we have used a Dynamic Linear Model to derive time-varying trends over 20-year periods in the Arctic ozone time series, measured in-situ by ozonesondes from 6 stations, from 1994 to 2022. The model accounts for seasonality, external forcing and 1st-order correlation in the residuals. As proxies for the external forcing, we have used tropopause pressure (replaced with Arctic Oscillation in the troposphere), eddy heat flux, the volume of polar stratospheric clouds multiplied by effective equivalent stratospheric chlorine, and solar radio flux at 10.7 cm for the 11-year solar cycle. Our results indicate that the ozone recovery in the lower Arctic stratosphere is not detectable. Though significant positive trends have been detected prior to 2017 at some stations, there are no statistically significant positive trends after 2017. Moreover, at a number of stations the trends after 2019 are rather negative and significant, varying between − 0.30 ± 0.25 and − 1.00 ± 0.85% per decade. Furthermore, the Arctic troposphere exhibited only statistically significant negative trends over 20-year periods ending in 2017 or later, varying between − 0.31 ± 0.27 and − 1.76 ± 0.41% per decade. These results highlight the importance of continued monitoring of the Arctic ozone.</jats:p>