Tag Archive for: article

Amazon Tells Ukraine Publication To Alter Its Article After It Links The Company To Ring’s Problematic Ukraine Branch

/in

An extremely-problematic wing of an extremely-problematic company is back in the news. Ring’s Ukraine division made headlines last fall when the presence of a “Head of Facial Recognition Tech” in the Ukraine office appeared to contradict Ring’s claims it was not interested in adding facial recognition to its cameras.

More disturbing news surfaced earlier this month, when it was discovered this office had allowed its employees to view Ring camera footage uploaded by users. Ring doesn’t just produce doorbell cameras. It also sells in-home cameras, making this revelation particularly worrying.

Beginning in 2016, according to one source, Ring provided its Ukraine-based research and development team virtually unfettered access to a folder on Amazon’s S3 cloud storage service that contained every video created by every Ring camera around the world. This would amount to an enormous list of highly sensitive files that could be easily browsed and viewed. Downloading and sharing these customer video files would have required little more than a click. The Information, which has aggressively covered Ring’s security lapses, reported on these practices last month.

Not only did the R&D team have complete access to customers’ recordings, so did Ring’s US-based engineers and executives. And who knows how many other people have accessed these recordings illicitly? When this access was granted to an untold number of Ring employees, Ring did not encrypt uploaded recordings. The company apparently felt encryption was too expensive to implement and would possibly limit revenue opportunities for the company as it aggressively moved into the home security market.

It turns out the company was using customers’ footage to train its AI to recognize faces and other objects. This would be the same facial recognition Ring swears it isn’t going to be implementing anytime soon.

Apparently, this abuse of trust has resulted in growth opportunities for Ring-Ukraine. A recent article by Ukranian publication Vector stated the office would be lending its expertise to other Amazon products, which possibly includes Rekognition, Amazon’s homegrown facial recognition program.

But that story was buried by Amazon PR shortly after it appeared, according to Sam Biddle of The Intercept.

I asked multiple Amazon representatives and Ring’s head of communications about the Vector article, including specifically what were the “many other Amazon projects” Ring’s Ukrainian staff now worked on.

Although Amazon ignored repeated requests for comment and Ring refused to discuss the subject on the record, it seems that the company did take action: Within hours of my inquiries, the text of the Vector piece was quietly edited to remove references to Amazon. Most notably, the entire quoted sentence about the “many other Amazon projects” the Kyiv office was working on was excised.

The author of the story told The Intercept he had nothing to do with the belated deletion. In fact, he was not aware of any editing until The Intercept brought it to his attention. An email from Vector’s editor-in-chief explained the situation, although not all that satisfactorily.

We published a news about rebranding, later pr-manager of Ring Ukraine called me and asked to take Amazon mention out from the article. Since I had a good relationship with manager, the article got just several dozens of views and I understood that everyone know that Ring is part of Amazon anyway, I didn’t even asked questions, said ok and took Amazon part out

It appears Amazon doesn’t want people to know it has given a problematic division even more responsibility. It also may be trying to head off another Ring-related PR nightmare by removing any text that might suggest Ring customer recordings are being used to train facial recognition software used by government agencies.

But it’s too late to change public perception. The scrubbing may keep Amazon from being linked to unfettered access to Ring camera recordings in search results, but there’s no separating Amazon from Ring. And there’s nothing here that suggests either company is moving away from leveraging user-generated content to fine-tune AI for the customers they really want: law enforcement agencies.

Permalink | Comments | Email This Story

Techdirt.

Air Force integrates missions, strengthens information warfare capabilities > US Air Force > Article Display – Air Force Link

/in

Air Force integrates missions, strengthens information warfare capabilities > US Air Force > Article Display  Air Force Link
“cyber warfare news” – read more

AFCYBER presents, participates in DEF CON 27 > US Air Force > Article Display – Robins Rev Up

/in

AFCYBER presents, participates in DEF CON 27 > US Air Force > Article Display  Robins Rev Up

Cyber warfare operators set up their computer stations before a capture-the-flag competition during DEF CON 27 Hacking Conference in Las Vegas, Aug.

“cyber warfare news” – read more

Swedish MEPs Announce Support For Article 13, Demonstrate Near Total Ignorance Of What It Actually Entails

/in

As MEPs get ready to vote on the EU Copyright Directive — and specific amendments concerning Articles 11 and 13 — many have not yet said how they are going to vote. However, two Swedish MEPs, Jytte Guteland and Marita Ulvskog, who many had believed would vote against the plan, have suddenly switched sides and say they plan to vote for it. In a rather astounding interview with reporter Emanuel Karlsten the MEPs reveal their near total ignorance of what Article 13 does and what it would require.

Guteland spoke to Karlsten by phone, and he asked all the right questions. It’s worth reading the entire conversation, but here are a few snippets with my commentary. When Karlsten pointed out the problems with filters, Guteland insisted that Article 13 doesn’t mean filters:

It is not a filter, it’s more about the sites taking reasonable steps, some of which can be free or cheap

That is… utter nonsense. The directive would require sites to block re-uploads of reportedly infringing material, and that means it requires a filter. There is no other way to do this. And there are no “free” filters. Currently, there aren’t even any “cheap” filters. Karlsten asks about this and Guteland changes the subject.

What do you mean when you say “free or cheap” ways to take the measures needed to stop copyrighted material from being distributed?

“You shouldn’t need to commit to expensive technical solutions, but measures taken must be proportionate and reasonable based on the content you have. If you’re a commercial player with huge amounts of content, then you’ll need to implement other solutions”.

Got that? You shouldn’t need filters, but… if you’re a company then obviously you can afford expensive filters.

From there, the interview gets even worse. Karlsten points out that lots of speech will certainly get taken down, especially since filters can’t determine what is parody or otherwise exempt, and Guteland’s response — I kid you not — is basically first “well, they can appeal their censorship” and when it’s pointed out that this could lead to content being censored for a long time, retorts that it’s somehow magically in the best interests of tech platforms not to censor the content too long:

There shouldn’t be lengthy court proceedings. Even big platforms have an interest in avoiding long court proceedings

“Shouldn’t be.” Apparently Guteland is unfamiliar with what has happened for years with notice-and-takedown regimes that are a lot less onerous than the ones that will be implemented post-Article 13. When pressed on this, she appears to give the Swedish equivalent of “Nerd harder, nerds.”

I see before me a mechanism being developed that doesn’t exist today, where now we have recognition technology, but no appeal process. In the future it should become second nature for platforms to examine whether content is satirical, so that it can be reposted quickly. That means it’s about recognition becoming a two-stage process where today it’s only one. That way it becomes easier to make judgments.

Full employment for satire-detectors! Also, the rest of this paragraph is utter nonsense. She acts as if there’s no current appeals process for content taken down today. There is and it’s a disaster that doesn’t work well at all. And under Article 13 it will be even worse, because the liability and penalties for leaving up the wrong content are much more severe than in the past. And that’s why she’s totally and completely wrong in saying that platforms will be quick to put this content up. Indeed, the only thing she’s right about is that they have incentives to “avoid long court proceedings.” And the way you do that is by KEEPING DOWN any content that might be questionable to avoid the liability.

It’s disappointing, especially as the vote is coming in just a few hours, that those supporting Article 13 still seem completely ignorant of how any of this works.

Permalink | Comments | Email This Story

Techdirt.