[Meai]

Ha that is interesting, but does that really apply to our fake news problem nowadays? Let's say a fake news site creates an article "Obama died in a hospital visit 3 weeks ago and was replaced by a robot."

Should we now make 150 different websites that spread 150 slightly different versions of this, who would gain from that? I'm just thinking out loud. Oh I think your point is that one of those 150 links would have to contain the truth and the rest of the 150 would slowly edge towards it. Kind of like this:

"Obama was injured and then replaced by a robot" "Obama was injured and then was given robotic implants to heal" "Obama was injured and given a pacemaker" "Obama visited the hospital for a routine checkup, minor cold revealed" "Obama did not visit the hospital 3 weeks ago, he was at a campaign rally"

Would you really say that you have helped the internet/humanity if you did that to every fake news link? Even if this is so sophisticated that it autogenerated new domains, new content... people would just revert to following CNN/Foxnews/[standard outlet]. Then the people who read these fakenews links will have an even harder time to figure out who to trust. Or is it maybe the goal to push people towards mainstream news outlets? I can only imagine that as a result of such an approach.

[rapnie]

I think the idea is, that when it is obvious that all information from unverified sources is false, then people will start to rely on the (cryptographically) signed, accredited sources when they want the 'real deal', and not let themselves to be misled when these signatures are missing. Some level of white noise is needed (enough to enounter multiple versions of the fake news) for people to recognize the value of checking the signature.

But maybe the white noise mechanism is not needed. There may currently be enough erosion of trust by 'black' noise to give platform builders the incentive to add the authenticity methods to their products and see widespread adoption of their use.

[Meai]

Foxnews proving to me that their latest article is actually from them only helps people who already trust that source. The reason why Assange etc. post signatures is because they dont have control over Twitter. The ownership of the domain already is a form of authentication/signature that is more than sufficient for just about everybody and source authentication definitely is not the main problem that fakenews is about. Verifiying that the author is who you think it is, is probably the smallest, most insignifant part of fakenews. Much more central is that the content isnt false. How do we prove that something is false? We usually can't, so we could at best try to find flaws in their thinking or quotes that are wrong and say 'probably false'. That's what fact checker sites are doing, they give out grades. In my opinion the approach of fact checker sites is the best we can do so far, the problem however is now identical to mainstream news: Corruption. These fact checkers inevitably mess up or maximize their grading to achieve goals for their ideology of purse, which has arguably already happened and now we dont trust factcheckers anymore either.

Maybe this is an uncomfortable thing to say but this entire escapade with fakenews may just be a natural cycle that happens when corruption becomes too much and competition is emerging. So if we accept that reasoning then fakenews is just one ugly side effect but there are also good side effects, like new news sites emerging which may use outrageous new content to get viewers or superior ethics as their selling point. Hopefully the latter prevails but the cycle of gaining / losing trust will continue for as long as human beings are fallible.

[politician]

> That's what fact checker sites are doing, they give out grades.

Hah! So also in Anatham, there are other machines that do this. The design the author wrote into the story involved two species of machines that work at full speed with 100% uptime to both revise and tweak the facts of a story and then, separately, to assign grades. Basically a world-wide generative adversarial network.

From the attacker's point-of-view, in order to deliver a false message they are forced to try to fight through a gauntlet of independent machinery that will first generate a bunch of alternatives and then will look at any particular story and assign a grade with knowledge that it's probably being attacked. That could be a very tough filter to consistently navigate, especially if our attacker is trying to conduct a broad campaign of misinformation.

From the victim's point-of-view, every piece of information they read now is associated with a score provided by their fact checking filters -- and there is no reason not to have multiple layers of grading filters.