It’s transliterated from Latin kalium, the first syllable.
To avoid confusion in symbols, please note k in Latin is /k/ in ipa and g in pinyin, voiceless but unaspirated. The pinyin k is voiceless and aspirated, which is ch in Latin and /kh/ in ipa.
Standard Mandarin no longer has the syllable /ka/, i.e. ga in pinyin. This is wrong, thanks for the comment 尬. But the syllable ga is irregular (see the last paragraph) and very infrequent.
甲 has initial 见 and historically pronounced as /kap/. From g to j (pinyin) is the second process of palatalization in Chinese. Some dialects didn’t go through this palatalization process, e.g. Cantonese, where 钾 is pronounced as gaap3. I came from a region where this process is not complete. We do pronounce 甲 as jia, but for other words with standard mandarin jia, especially colloquial ones, we pronounce them as ga. So my first instinct at seeing the Latin ka is to associate it with ga in my local dialect, and then immediately translate it to jia in standard mandarin. The initial translator probably went through the same mental process.
You can see here https://blog.sciencenet.cn/blog-612874-1346630.html and related blog for a full list of the origins of Chinese element names.
My opinion: The problem with 尬 is most likely that its sound is irregular and thus ambiguous if used as a phonetic component. Its induced Standard Mandarin sound should be jie4. I’m not sure where does ga4 come from, but mostly it came from a colloquial layer in some southern dialect. It's from Wu dialect (thanks to @alephalpha for the comment). Probably also because it’s composed of left and right parts. If we add an 金 to the left it seems a bit wide to write. Also probably because the syllable ga is very infrequent in standard Mandarin; it didn’t even come up to me at first. Also kalium has a i in the second syllable, making it more likely to be associated with a palatalized sound.
