"Stable" here means mechanically stable (they don't fall apart as quickly as was expected), not radiologically stable (the radioactive elements, of course, still decay at their normal rate).
For radionuclides more long-livd means less radioactive per unit of time, though. Of course uranium compounds can be toxic in the chemical sense as well, but I’m unsure if uranium compounds make it to the top threats in that list compared to e.g. arsenic and lead.
Of course radioactivity needs to be respected, and the exclusion zone is there for a reason due to factors such as hot spots.
This is "long-lived" in the ecological sense, not radiological. The grains of uranium oxide dust remain grains of uranium oxide dust, rather than breaking apart. If you ingest a few molecules of uranium oxide, it pretty much doesn't matter, you're fine¹. Ingest these dust particles, not so much. It's a question of concentration, and it seems that it's not diluting out naturally.
(P.S.: the radiological stability and lifetime of uranium doesn't make much sense to question; the dust flakes aren't large & concentrated enough to significantly shorten their half-life due to their own neutron emissions cascading and this isn't what the study was researching. Note the article talks about weathering: "It remains largely unclear why these particles weather at different rates in their environment.")
Of course radioactivity needs to be respected, and the exclusion zone is there for a reason due to factors such as hot spots.
¹ humans contain, on average, 90µg of uranium. [https://www.iaea.org/sites/default/files/DU_Eng.pdf]
(P.S.: the radiological stability and lifetime of uranium doesn't make much sense to question; the dust flakes aren't large & concentrated enough to significantly shorten their half-life due to their own neutron emissions cascading and this isn't what the study was researching. Note the article talks about weathering: "It remains largely unclear why these particles weather at different rates in their environment.")