> The exposure draft doesn’t give Google/Facebook the right to opt out by “discriminating” against (ie. dropping) protected news sources.
It doesn't permit digital service corporations from discriminating against protected news sources on account of their participation in the system. It does seem to permit actions that affect all news sources equally.
However, more frustrating and worrisome to me is that the exposure draft doesn't seem to define what conduct by Google/etc triggers these provisions. Some clauses apply to news content "made available" by the platform, but there's no definition in this bill of what "made available" means.
That could be read to require some sort of hosting of the content (e.g., rehosting an article or using a long extract), or it could be read to include mere linking. That's a huge difference: Hacker News could be a 'digital platform service' under the latter definition, but not the former.
How will they determine what counts as "news"? Are there be significant penalties if whatever algorithm they come up with to decide what counts as "news" has any false negatives? If so, that might make "stop showing news" an unworkable approach.
Yes, I think the last line in the anti-discrimination section is there specifically to thwart that, but one can still easily imagine a lot of legal acrobatics to get out of that.
As previously stated above, that method is explicitly disallowed by the law:
> Responsible digital platform corporations may not discriminate between the news businesses participating in the code, or between participants and non-participants, because of their participation in the code.
I don't get what you're trying to say. If Google stops showing news from participants, but show news from non-participants (due to false negatives causing their algorithm to fail to identify those sites as news), would that not be obvious discrimination according to the above quote?
That just comes back to my original comment then. How do you determine what counts as "all news"?
You can't just use the list of "businesses participating in the code" because that's considered discriminatory and illegal under the new law. And if you decide using any other method, you risk accidentally classifying one of the "businesses participating in the code" as "not news" and showing it in search results anyway, which is also illegal under this law.
The best solution I can think of would be to use the list of businesses participating in the code as a starting point, and then add other non-participants through some other method. But then you run into a similar problem: what happens if you accidentally classify a non-participant as "not news" and they see a surge of traffic because they're one of the only news organizations in Australia that shows up in Google search results? Wouldn't that also be considered discriminatory and illegal under the new law, since that organization wouldn't have shown up in the results if they were a participant?
I'm fairly sure that Google already has algorithmic classification for "news" that's accurate enough for these purposes. If not, "is this a news article" is quite amenable to machine learning techniqes. All that's left is ensuring that registered news sources are definitely classified as news.
It doesn't permit digital service corporations from discriminating against protected news sources on account of their participation in the system. It does seem to permit actions that affect all news sources equally.
However, more frustrating and worrisome to me is that the exposure draft doesn't seem to define what conduct by Google/etc triggers these provisions. Some clauses apply to news content "made available" by the platform, but there's no definition in this bill of what "made available" means.
That could be read to require some sort of hosting of the content (e.g., rehosting an article or using a long extract), or it could be read to include mere linking. That's a huge difference: Hacker News could be a 'digital platform service' under the latter definition, but not the former.