Agreed. I think, as written, it's extremely difficult to read as anything but a straw man of itself. I had to pause several times and steel man the concepts, because as written they were just so obviously a bad idea that I couldn't picture you having possibly intended it that way.

Specific examples of my confusion:

• It wasn't clear to me until halfway through the article who this contract would be between. (Papers and readers, I assume?) I initially interpreted it as a contract between, say, a politician and someone wishing to interview the politician. Widespread use and acceptance of this sort of agreement seems like it would be a net negative. There do seem to be a few positive use cases though. Imagine Peter Singer agreeing to be interviewed about the trolley problem or something, on the condition that the interviewer sign an agreement. Such an agreement might prevent the interviewer from publishing the interview unless Singer reviewed the article before hand and said ok. This would limit the temptation to sensationalize. Perhaps these agreements should be made on a per article basis, so it’s still possible to publish negative things on politics.

• That brings me to the phrase “no negative press agreements”. This sounds like an agreement not to publish anything with negative emotional valence. No bad news. No scandals. No reporting on wars or earthquakes, and maybe not even bad weather. Obviously this isn’t what you mean. I think you are saying not to report on shark attacks, cop shows, and things that make people terrified for no statistically good reason. Or perhaps not to publish any story which has negative expected utility according to some utilitarian framework. If either of these are what you mean, perhaps a better term would be “fear-mongering free journalism” or “no hit-piece agreements” or “no shock-piece agreement” or something. I don’t really like any of those, but I’d have to think about a better name for it for a while. It’d be nice to have an intuitive name that conveys the approximate meaning succinctly. If that’s not possible, maybe a deliberately nonobvious phrase, ~~so people don’t mistakenly think they understand, and unwittingly strawman the idea~~ . (See the purple ball thing. Also, apparently I’m really bad at doing that. EDIT: Also, apparently Strikethrough syntax in markdown doesn't work on LW. So much for that joke.)

• It would also be helpful to start the piece with a statement of the problem you are trying to solve with this. Perhaps a couple examples of “negative” journalism and a couple “positive” examples, to illustrate what the differences are which you are trying to highlight. Then, try and construct a precise rule of phrase which precisely divides the two groups with as little ambiguity as possible. The more objective this rule is, the better, but that’s a difficult task. The thinking would probably take much longer than 45 min, though, unless you've already done most of it.

Please don’t take any of this as discouragement, though. I think you have a core idea in there which might be useful. I’ve given some thought to this sort of thing, but as a browser add-on that acted like an add-blocker, rather than as a contract with news agencies. This might be further inside the Overton window, too.

Also, perhaps you/we could ask the mods of a couple big subreddits to try various forms of the rule out for a month, by filtering out various definitions of “negative” news. Some already approximate this unintentionally, by disallowing sensational titles or certain “low quality” content. This should give a little empirical data on what works and what doesn’t, before anyone suggests this to any news outlets. It’s really hard to get entrenched powers to change, let alone change twice.