The Link: Microsoft is teaming up with Semafor on AI-assisted news stories [Engadget]
The tl;dr: Semafor, with Microsoft's help, will create a Signals feed with AI-assisted but journalist-written news. The money quote: Specifically, Semafor's team will use AI tools to rapidly find breaking event reporting from other news sources around the world in multiple languages, while providing translation tools. An article might therefore include Chinese, Indian or other sources, with reporters adding context and summarizing the different viewpoints.
The good: The fact that AI text generation is prone to conceptual errors that can render useless (or worse) even the most linguistically flawless article, so you need humans there, is key, and it's good that they are designing the process with that in mind.
The risk: Deskilling. Surfacing breaking news, "summarizing viewpoints," and providing context are important activities, but at its core, I believe, the unique professional skill and social value provided by journalists (what makes news media something more than link farms with legacy brands) is generating new news through research, analysis, and being nosy both online and offline. The more you focus new technologies and investment on the first set of activities, the more likely it is you're underinvesting on the latter. (I don't mean only data journalism, which is important but only a piece of it.)
The strategically tricky: Surfacing breaking news and commenting on them is more or less what social networks do — at this point they do that more than social networking in a more literal sense. A professional version of that with solid standards is welcome, but overall it's not something you want to bet the future of journalism on.
The opportunity: Everybody — journalists, news media, non-news media, good actors, bad actors, non-profits, etc. — is trying to figure out what's next with news, sometimes in the sense of "how do we save this socially necessary function," sometimes in the sense of "how much money can I get off this brand before I finish killing it." Times are ripe for trying out wild things! Being the radical optimist that I am (no, really) I think the most interesting kind of wild thing to think about is to start with the social function of journalism (which in my head isn't too far from "finding out and making public whatever organizations and people with power would rather not") and move on to "what sort of toolset, processes, etc, best leverage our full set of technologies and resources to do it better?" It's an important question and nobody seems close to having an answer, and those are some of the best problems to work on.