The criteria for resolving yes is that the Arctic must be fully ice-free at any point in time between now and the start of 2050. The Arctic does not need to be fully gone by the year 2050, it just needs to be ice-free at a moment in time.

Whether the Arctic is fully ice-free will be based on news articles from reputable sources. If there's any controversy, I will discuss about it with Manifold users in the comments.

As a last resort of resolution criteria, this question will resolve Yes when the Arctic ice coverage (at a threshold 15% ice density) falls below 100,000 km squared, as reported in numbers from http://nsidc.org/arcticseaicenews/charctic-interactive-sea-ice-graph/

The Arctic has less ice cover in summer than in winter, so it is plausible that in summer, the Arctic will be fully ice-free, but in winter, it will reappear. This type of scenario will resolve as Yes.

Relevant info:
https://english.elpais.com/science-tech/2023-06-07/the-arctic-is-set-to-completely-melt-for-the-first-time-within-two-decades.html
A study supported by observations from NASA and ESA satellites and a sophisticated climate model predicts that the first ice-free September will arrive between 2030 and 2050.

https://phys.org/news/2019-11-arctic-ocean-ice-free-year.html
According to a new study by UCLA climate scientists, human-caused climate change is on track to make the Arctic Ocean functionally ice-free for part of each year starting sometime between 2044 and 2067.