Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×

Submission + - SPAM: Algorithm automatically spots "face swaps" in videos

yagoda writes: Andreas Rossler at the Technical University of Munich in Germany and colleagues [] developed a deep-learning system that can automatically spot face-swap videos. The new technique could help identify forged videos as they are posted to the web.

But the work also has sting in the tail. The same deep-learning technique that can spot face-swap videos can also be used to improve the quality of face swaps in the first place—and that could make them harder to detect.

The new technique relies on a deep-learning algorithm that Rossler and co have trained to spot face swaps. These algorithms can only learn from huge annotated data sets of good examples, which simply have not existed until now.

Link to Original Source

Submission + - ICANN global Whois system expected to shatter on May 25 with new GDPR law (theregister.co.uk)

monkeyzoo writes: In a letter sent to DNS overseer ICANN, Europe's data protection authorities have effectively killed off the current service, noting that it breaks the law and so will be illegal come 25 May, when GDPR comes into force.

ICANN now has a little over a month to come up with a replacement to the decades-old service that covers millions of domain names and lists the personal contact details of domain registrants, including their name, email and telephone number.

ICANN has already acknowledged it has no chance of doing so. The company warns that without being granted a special temporary exemption from the law, the system will fracture, perhaps even resulting in the Whois service being turned off completely while a replacement was developed.

Critics point out that ICANN has largely brought these problems on itself, having ignored official warnings from the Article 29 Working Party for nearly a decade, and only taking the GDPR requirements seriously six months ago when there has been a clear two-year lead time.

European agencies responded and tore ICANN's plan to shreds, pointing out that it needs to be much more precise and to include both compliance and auditing functions. Critically, however, it did not address ICANN's request for a moratorium.

Even the idea of a moratorium appears to have been invented by ICANN. This is no evidence of a similar request from any other industry, and the GDPR is, after all, a globally applicable law that affects everyone.

Submission + - SPAM: CFPB left our private data more vulnerable than ever

schwit1 writes: Without your knowledge or permission, Consumer Financial Protection Bureau(CFPB) collected and warehoused your most private bank records and continued to sweep them up — despite repeated warnings the data wasn’t being properly protected. Now there’s a good chance your personal information could be in the hands of identity thieves or even terrorists.

The government isn’t sure who has your information. It only knows the CFPB databases have been breached by outsider threats potentially 1,000-plus times. That’s according to a recent investigation of cyber-intrusions at the Consumer Financial Protection Bureau, where the sensitive information is stored.

The number of confirmed breaches of consumers’ personally identifiable information is “just north of 200,” revealed Mick Mulvaney, the White House budget chief who took control of the CFPB late last year, in testimony to Congress. “We think there’s another 800 [incidents of hacked information] that we suspect might have been lost, but we haven’t been able to nail that down.”

The agency also shared this data with outside agencies and contractors, including state attorneys general, trial lawyers and civil-rights organizations interested in filing class-action lawsuits against banks, according to regulatory documents and congressional testimony.

Link to Original Source

Submission + - Doctors tried to lower $148K cancer drug cost; makers triple price of pill (arstechnica.com) 2

Applehu Akbar writes: Imbruvica, a compound that treats white blood cell cancers, has until now been a bargain at $148,000 per year. Until now, doctors have been able to optimize dosage for each patient by prescribing up to four small-dose pills of it per day.

But after results from a recent small pilot trial indicated that smaller doses would for most patients work as well as the large ones, its manufacturer, Janssen and Pharmacyclics, has decided on the basis of the doctors' interest in smaller dosages to reprice all sizes of the drug to the price of the largest size. This has the effect of tripling the price for patients, and doctors have now put off any plans for further testing of lower dosages.

Submission + - Why does copy and paste copy styles by default?

Arnold Reinhold writes: Why does the paste function on so many software packages apply the style of the source document rather than the style at the insertion point? Getting the later behavior, which is what we want 95% of the time, requires complex key combinations or going to the menu and selecting “Paste special.” Except for composing ransom notes, how often does anyone want to drag over a different style along with a chunk of text they are copying?

Comment Re:Zuckerberg knew his questioners lacked knowledg (Score 1) 235

They were certainly briefed on the issues, but I didn't get the impression that many--if any--had much real understanding of the topic. Especially when it comes to the ramifications of AI working with the data.

On the subject of third party icons, I was specifically referring not to the 'like' buttons, but those offering to log-in new users via a Facebook link to their identities. I expect that these transmit a lot more data in both directions. (Then again, I certainly could be wrong. The 'like' button may do more than I have assumed.) It would be very interesting to know if any /. folks have examined the operation of these things and what was found.

Comment Re:Facebook has done nothing illegal (Score 1) 235

"These people's ignorance is not a reason to blame Facebook for anything."

Yes it is! Facebook victimizes people who don't understand the technology. It's not realistic to believe that everyone even has the capability of understand this kind of thing, especially since there's precious little mainstream discussion. There are a lot of technical people who don't understand the issues. I will admit that there's some precedence that's been set in holding people responsible for understanding legal jargon, but I believe it only stands because lawyers make the laws. A lot of what's done on this basis is morally reprehensible, and the same is true of many Facebook activities. Would you defend payday lenders?

AI is not yet smarter than people in every respect, but it certainly excels in producing results using huge data sets. No human could do that.

Can I interest you in a game of Go?

Comment Re:Facebook has done nothing illegal (Score 4, Insightful) 235

"The information they keep about people was given to them voluntarily — either by users themselves, or by their friends and acquaintances. And what they now know, they are free to share — sell, give away, publicize, it is up to them."

I disagree. Most of the people who provided the data to Facebook had no idea that it could be used in the way it is. Most Facebook users still don't know what's being done with their data, and that's exactly the way Facebook wants it. To say that these people willingly handed it over is like saying people scammed during the savings and loan scandals should have known better. When is the last time you read three or four pages of fine print legalese before signing up to use a website, or when applying for a mortgage? Even if you're determined to read it, you won't fully understand it unless you're a lawyer.

And when they scrape copies of every text message you sent with your cell phone, back when they could do it without asking; when they collected the names and phone numbers of all of your contacts; when they kept logs of who and when you call; when they keep logs of where you've been day after day out using location data from your phone--when they store all of this information, combined it with data from your Facebook profile, then put AI engines to work on it--I would say that those actions were all illegal invasions of privacy. Zuckerberg and friends should be tried for illegal eavesdropping.

And when that's done, they should start on Google and Microsoft . . .

Some people worry about what will happen when computers get smarter than people. They don't realize that it's already happened. AI can process huge volumes of data that humans could never hope to handle. The insights/information that can be derived from the kind of data that Facebook and Google keep boggles the mind. But don't believe me. Google it for yourself and you'll see. Maybe start with 'psycho-analytics'.

Comment Zuckerberg knew his questioners lacked knowledge (Score 5, Insightful) 235

Zuckerberg took maximum advantage of the fact that the questions came from people mostly lacking the technical knowledge to judge his responses. For example, when asked if Facebook could track users across devices, he acted as though he didn't know. Is there anyone here who believes that? I wish we could ask him a few questions on Slashdot!

Zuckerberg also said that Facebook doesn't share user data, just uses it to predict which advertisements users are likely to respond to. In that case I'd really like to see what gets sent when someone uses Facebook to sign into a third party website.

Submission + - Facebook sent doctor on secret mission to ask hospitals to share patient data (cnbc.com)

NichardRixon writes: According to a story by Christina Farr on the CNBC website,
" Facebook was in talks with top hospitals and other medical groups as recently as last month about a proposal to share data about the social networks of their most vulnerable patients.
  " The idea was to build profiles of people that included their medical conditions, information that health systems have, as well as social and economic factors gleaned from Facebook.
  " Facebook said the project is on hiatus so it can focus on "other important work, including doing a better job of protecting people's data."

Comment Re:Isn't This What Facebook Was Engineered To Do ? (Score 1) 119

Of course you're technically right, there's no getting around the fact that the whole discussion is by definition political. But what I mean is that it doesn't need to go the route of us vs them / red vs blue. In my opinion it's not a partisan issue, it effects all of us, and no ones hands are clean. My hope was that we can all see that.

Comment Re:Isn't This What Facebook Was Engineered To Do ? (Score 1) 119

"In short, if Hillary had won, nobody would know or care about this. Clearly this is political.

"Am I wrong about that? "

I hope so. I wouldn't put it past either political party to misuse the Facebook data, similar, or just about any other dirty trick. It seems to me that the main reason this has become a disagreement between right and left wingers is because both parties would rather divert attention from the erosion of Constitutional rights that is happening so rapidly in this country. Personally, I'm more concerned that Facebook or Google may know more about me than I know about myself then I am over which political party got caught in a dirty trick.

That isn't to say that I don't prefer one over another because of the "lesser of two evils" philosophy, but would rather not bring that into the discussion.

Thanks for the reasoned reply.

Slashdot Top Deals

If you think the system is working, ask someone who's waiting for a prompt.

Working...