Multipolar and multilingual categories, such as 'World' or 'Europe', should consider non-English sources

JR

Hi. As far as I can tell, the multipolar/multilingual categories, such as World/Europe, only appear to consider English sources. Having followed these two for months now, I have never seen a non-English source cited.

In my view, this serves to elevate English-language sources when it comes to these topics, which brings in the sources' own biases regarding topic choice and coverage, which sadly propagate to Kagi News itself.

This is a bit of a wasted opportunity, in my opinion. With Kagi's strong LLM synthesis and translation capabilities, I think there is an opportunity to offer truly global 'World' and 'Europe' news that transcends the language boundaries that have separated us for so long.

The engine/LLM that produces the multipolar/multilingual categories, such as 'World' and 'Europe,' should consider all reputable sources from these regions, regardless of the source language. This would improve the depth and strength of coverage and reduce Anglophone biases in the news selection. It would offer alternate perspectives and views and elevate local sources of global news, which are likely to have more detailed coverage of global news happening in their backyard.
In fact, broad multilingual coverage that is in agreement should be one of the strongest indicators that a story is worth reporting.

I think it should work the same way these categories work now, just with more diverse sources.

Perhaps you can consider rolling it out as an experiment so that the quality can be ascertained and users can consider both the English-only version of these categories and multilingual versions of these categories.

EDIT: This should also be done for interest domain categories such as 'Gaming'.

westerntragopan

does the settings not help? just wondering as English is the only language I speak.

JR

westerntragopan I cannot test this, as the only non-English language I speak confidently is not offered, but to my knowledge, this is not how it works.

From what I can tell, the categories consume RSS feeds from relevant sources. Where these sources refer to the same news item, a consensus report is synthesised in the source’s original language. For national categories, this is in the local language, but for global, multilingual, or specific-interest categories, this is done in English, based on English-language sources. While English-language sources aren’t always from Anglophone countries, the vast majority are, which skews the coverage toward Anglophone media landscapes.

These original reports in the original language are then translated into every target language. If I switch my language to another one, the sources used are still English-speaking. Even when a non-English-speaking country’s source is used, such as Le Monde (https://www.lemonde.fr), the newspaper’s English-language coverage is used, not its bread-and-butter French content.

For me, one of the major value propositions of Kagi News is the ability to read news from countries whose languages I don’t speak. Of course, this was always possible by manually searching and translating, but Kagi News synthesis lets me see a high-level view of what the country’s media is discussing, which then helps me decide which coverage I should actually translate and read. To me, this is invaluable, and not including non-English sources in multipolar or multilingual categories is a missed opportunity.