Yes, they adapt
A recent large American study compared weed control strategies based solely on glyphosate with those using other weed control methods prior to crop emergence (https://academic.oup.com/pnasnexus/article/2/12/pgad338/7457920). The result was clear. The more and more glyphosate is used, the faster the weeds adapt. MirageNews also reports (https://www.miragenews.com/glyphosates-weed-control-efficacy-declines-over-1137489/  ).

 How do the plants adapt?
In some cases there are mutations that detoxify the herbicide and render it ineffective. Only mixtures with additional herbicides remain.
In other cases, the entire plant changes in such a way that, for example, less herbicide is absorbed in the leaves and transported to the roots. In this case, an ever-increasing dose still helps against the weeds, but not the environment or the costs.
Selection for evasive behavior is similarly effective but even more exciting. For example, if glyphosate is always sprayed in an region on a certain date, the plant germinates so much earlier that the normal glyphosate dose no longer kills the plant, as it is already too large. The dose must therefore be increased. Or the plants germinate later so that they are no longer affected by the first spray. In this case, the crop has to be treated more often to maintain the effect, which puts an increasing strain on both the wallet and the environment.
The already visible consequence: according to study results, the efficiency of glyphosate decreased by up to 31.6 % within 10 years when glyphosate was used alone (Conyza canadensis, Canadian horseweed). There will soon be no effect at all.

Does KI help?
According to the authors, artificial intelligence and partial area or spot treatments do not help against all these efficacy losses. Although the technology reduce the current pure spraying agent costs, it leads to the same pressure to adapt and the same loss of efficiency – just a little more efficiently with  less actual glyphosate use.

What helps?
Only when different herbicide strategies were used in the same year was such a reduction not observed.
Glyphosate is therefore not the “silver bullet” cure for everything and forever. Even its efficiency in areas that are more difficult to replace will only be sustainable in the long term if completely different weed control measures are used regularly and wherever possible. To this end, the authors of the study clearly recommend all non-chemical strategies. In order to make a better selection, these must be tested for their main beneficial properties and undesirable side effects and used accordingly.