Hacker News new | more | comments | ask | show | jobs | submit login
Coming Soon to a Police Station Near You: The DNA ‘Magic Box’ (nytimes.com)
98 points by SQL2219 29 days ago | hide | past | web | favorite | 58 comments

This is really worrying.

> To illustrate the point, he selected a sample from a 52-year-old Bensalem resident who had been pulled over the previous day for running a red light.

They're doing DNA matching against people pulled over for misdemeanors (speeding and an outstanding warrant for "retail theft"). And this is apparently now being connected to a national database.

> To collect DNA, police in Pennsylvania must obtain consent from people under arrest. Ninety percent of those asked say yes […] Asked why so many people would consent to give DNA, he said: “I have no idea. But criminals do stupid things.”

I know why. It's because you've got someone in a position of power asking those they have power over for something they don't really understand. People will say yes because they hope that agreeing to the demands of their captors means they'll be treated better.

Further in the article

> “We’ll say, ‘Listen, we’ve had stuff in the area. Would you mind giving us consent to take your DNA, so we can rule you out for committing any crimes?’”

The way they're phrasing this makes it even worse. They're clearly not explaining what this actually means, they're pretending like this is to help the suspect, and the suspect probably has no idea that this is being compared to a large database of crime samples.

> Mr. Harran called this criticism “total nonsense.” His officers do not target particular groups for DNA collection, he said: “You have nothing to fear if you’re not going to be a criminal.”

This is the exact same argument made against the right to privacy.

Yeah, totally ignorant. If there is a database with DNA from everyone, you can trivially frame anybody.

It is usually easy to obtain DNA from a specific person and put it at the crime scene.

I'm sure Mr Harran's DNA won't be found in any database.

My understanding that these databases effectively contain a checksum which has a very low chance of collision. I don't know if you could actually make DNA material from that checksum.

The comment you replied to did not mention making DNA material. DNA can be obtained directly from the target.

CCC members has obtained fingerprints from politicians, for example. DNA is no harder.

Or don't target a specific person, but place random hair found in a public toilet.

Not only is obtaining DNA contaminated samples not harder than fingerprints, it is orders of magnitude easier.

I can believe they don't target "particular groups", just "every group".

Anybody who parrots ‘nothing to hide’ or anything similar needs to read The Trial. If we’re heading towards a dystopia, I don’t think it will be Orwell’s, I think it will be Kafka’s.

Anyone who parrots 'nothing to hide' should have a public webcam in their shower.

"My problem with quips like these -- as right as they are -- is that they accept the premise that privacy is about hiding a wrong."


suggesting a webcam in the shower is not about "hiding a wrong".

Parent was agreeing with the grandparent.

While that argument is valid, it doesn’t touch on the true nefariousness of ‘nothing to hide’. Which is how a secretive bureaucracy can interpret isolated facts any way they please, and place the burden of proof onto the accused. A real life example would be the no fly list. How do you get on it? How do you get off it? The best example in fiction would be Kafka’s The Trial, imo.

Weird, because this genuinely wouldn't bother me, while being on a DNA database would.

I don't understand your point. It's possible to not want your nude photos public while also being comfortable submitting your DNA to a government database.

I personally probably wouldn't submit my DNA without a particular reason, but I don't see what's the contradiction here?

Ever been in a liason with a policemen wife and had your DNA turn up in a murder case?

You're missing the point. Everyone in this thread agrees that submitting DNA voluntarily is a bad thing. There is no need to pursue that part.

The confusion felt by some in this sub-thread is the analogy between 'having nothing to hide' and 'having a pub webcam in your shower'.

I don't see the connection either.

I'm not sure. I have always interpreted the "Nothing to hide" line as being past tense. Mr. Harran's statement is arguably future tense.

> “We’ll say, ‘Listen, we’ve had stuff in the area. Would you mind giving us consent to take your DNA, so we can rule you out for committing any crimes?’”

But this is exactly what they are doing? It is an accurate description. Of course they keep it for future crimes too.

It's actually the inverse: They suspect the individual but have no other recourse but to suggest to the suspect that they want to rule him out.

It's akin to when police ask, "Do you know why I pulled you over?" The question is open-ended and if the driver admits to any other wrong-doing, other than the reason for the stop, that's now introduces reasonable suspicion (by the act of admission) to other crimes.

Is it illegal? Technically, no. (Unless you go down the entrapment road.) Is it morally wrong? I think most people would agree that it is because the consequences aren't told the individual. (The informed consent argument.)

Also, Gattaca.

> And this is apparently now being connected to a national database.

And yet we still don’t get a national id. This is the worst of all worlds outcome.

This doesn't just effect people who willingly give up their privacy. Once a majority of our citizens are sampled, it'll be easy to paint the lack of a DNA sample for someone as a suspicious outlier. At that time, the public won't see the harm in mandating collection for everyone.

Also, consider how DNA evidence is used. If you've got a 1-in-a-million chance of being matched, the hash might find you as a suspect for a crime. Then, during the trial, the prosecutor will tell the jury there's a 1-in-a-million chance they got the wrong person since the DNA matched. Never mind that the DNA is no longer statistically independent evidence (they used it to find you, of course it matches). If this approach is successfully challenged because it's not independent evidence, then, the standard practice will bring in "parallel construction" techniques. Once you're identified, prosecutors will use other evidence that doesn't have to meet "beyond a reasonable doubt" as the official reason they found you. Then, they'll use DNA in the trial as if it were independent evidence.

  Once a majority of our citizens are sampled,
  [...] the public won't see the harm in
  mandating collection for everyone.
The government doesn't need a DNA sample for Bob if they've got samples for Bob's family.

According to [1] the Golden State Killer was found by matching a relative in a public DNA database. Although obviously if the DNA match says 'sibling' or 'parent' there are fewer suspects to comb through than if the match says 'third cousin'.

[1] https://www.vox.com/science-and-health/2018/10/12/17957268/s...

If it's a 1-in-a-million chance of matching a random person, and there are ~7000 million people on the planet, on average this method will finger 7000 people for each crime?

Yeah, the 1-in-a-million number isn't quite right, but there are other reasons why DNA evidence isn't perfect. Twins and chimeras for a start. And then you go on to washing machines, which are great for spreading DNA from multiple sources all over lots of things that tend to get analysed in a forensics lab.

The rare instances where police plant drugs on suspects likely happens more often than it's caught, and planting DNA at a crime scene or asserting they got their sample from the scene is only marginally more trivial.

How is an investigator asserting that they found the trace DNA from hair or saliva they are presenting as evidence at the crime scene any different from any other assertion?

While these are exceptions, horrible abuses have already been committed by lying forensic specialists. In Canada, we had a forensic lab fabricate drug use evidence against mothers, who ended up losing their children. http://projects.thestar.com/motherisk/ , similarly, a pathologist was charged for fabricating evidence as well, https://en.wikipedia.org/wiki/Charles_Smith_(pathologist) .

These are low likelihood scenarios, but with an extreme impact that has consequences for the integrity of the system. These are in recent memory, and only examples who were caught.

This is the vector that will be abused the most. There is no way to verify, other than the officer's word, where the DNA sample came from. Plus the fact that we shed DNA constantly, littering our own DNA any place we visit makes random DNA found in a location really just random DNA in the majority of cases.

Might as well call it the data leak jackpot. You can change your password, but you can't change your DNA.

> Unlike DNA labs, Rapid DNA machines do not have rigorous protocols governing the handling of samples.

> “There really are no actual rules written anywhere,” Detective Vandegrift said.

> With a few clicks, Detective Vandegrift uploaded it to the county database.

> In contrast, county DNA databases are unregulated. In Bucks County, the DNA database has begun to include genetic material from people whom police consider “even just a suspicious subject,” Detective Vandegrift said. Mr. Harran called such cases “one of the greatest uses of this instrument.”

Instead of a movie about changing your grades, now it will be changing a DNA file and getting them convicted of a crime.

It gets even worse. The way "DNA matches" occur in the judicial system is by looking at a set of SNPs, not whole genome comparisons (obviously). So all you need to do it is take the SNP set, the leaked SNP set, synthesize oligos matching the person you want to frame, then sprinkle the oligo collection at the crime scene...Guess what shows up in the PCR? You do.

I thought STRs were used for DNA fingerprinting? Has that changed with the advent of cheap SNP chips?

Looking at the 'rapid DNA machines' on the market, looks like it is still STR.

Someone should sell them on the idea that they can use SNP to generate an approximate photofit

I don't understand a thing. Where can I learn more about this threat?

Hmm, learn about polymerase chain reaction first then learn about DNA synthesis. That should get you close.

I find this story disturbing & the implications are frightening (many are addressed in the article).

But one thing strikes me as REALLY worrying: they call it the "MAGIC" box.

IMO, if there is ONE thing that you DO NOT WANT in serious such matters, it is the propagation of the -common- belief that science/technology == MAGIC.

This is already a problem with computers, as people accept plenty of nonsense "because the computer said so", and to be fair technologies like computers have become so complex, it's difficult to NOT have a "magic" feeling to some extent (let's not get into sciences which are IMO even worse).

Relying blindly on something that you don't really understand (its limits, its strengths & weaknesses, the context in which it should be used), is a RECIPE FOR DISASTER.

I have a tongue-in-cheek joke about this: "Data doesn't lie."

The problem with most people is that they agree with that notion and never second-guess it. ...but anyone who's ever modified anything in code or a database or the like will know that to not be the case.

Considering it "magic" also proves the following posit: The result of the use of that technology is potentially as faulty as the operator whom doesn't understand it.

Case in point: "Tides go in, tides go out. You can't explain that."

Perhaps now the widespread rape kit backlog can be analyzed?

I've often wondered how much of the backlog was self-imposed. As in, the patsy might actually go free if the tests were completed.

I don't know how much it's self-imposed, but I imagine if 100% of the samples were tested and the evidence reanalysed then we'd be seeing a mass exodus from the prisons.

Don't hold your breath that lowered cost would improve anything... https://medium.com/in-justice-today/five-ways-the-media-driv...

Local police departments already suffer from a lack of transparency and oversight, especially regarding evidence collection. This just seems like a nightmare waiting to happen in too many ways.

Meanwhile - Unsolved Shootings Are Rising:



I have no issue with collecting DNA from those convicted and sentence of a major crime. I do have serious issues with collecting from people for traffic stops or similar. If anything I think it should be unlawful to ask for such consent without a representation present and it should in no means be considered probably cause to arrest or detain if consent is not given. Simply put the general public is not versed enough in the law to give a good response.

I could see a system whereby you give a one time use consent to clear you of suspicion, requiring the authorities to destroy the sample and record after it fails to match. Even then I would suggest a court order being required to force anyone to provide a sample and not providing a sample does not give them probable cause or the like.

The sad part is far too many are willing to hand over much of their life and responsibilities to government in return for free stuff and are willing to see you are forced to give up the same. be wary of any politician who tries to frame this as moral duty or requirement because some will

>I have no issue with collecting DNA from those convicted and sentence of a major crime.

While the justice sysem still has gross lapses, I do mind - for the most part. A guy who gets released after 20 years, because he was proven innocent, is still going to be in the database.

There are no protections for the accused in that scenario. So, let's assume that the manufacturing of a specific genome sequence becomes much easier and much cheaper. What's to prevent a police office, with a justice boner and who infallibly believes that the guy is guilty - despite evidence to the contrary, from abusing the information and "discovering" this individual's DNA at another crime scene?

While it might seem like a far-fetched idea, there's been a history (though it may be seldom) of people abusing the "solid science" position to frame people.

There's another comment[0] in this thread that demonstrates that this is not only plausible but it has happened.

[0] - https://news.ycombinator.com/item?id=18997702

>“You have nothing to fear if you’re not going to be a criminal.

...and there it is: The rationale that anything is worthwhile, as long as they're catching the bad guys.

Nevermind that "being suspicious" is warrant enough to get you dropped into a database, for life - which is something they don't explain, when they're using the "we just want to rule you out" narrative.

I've read the comments here and I have nothing to add other than it's sickening that the police see nothing wrong with this.

I guess when you're on the side of the blue line that is protected from the excesses of law enforcement you forget about what the threat it poses to the rest of us.

It will be during our time that we apologize to the younger generation for the world we created

Dr. Lamar: Congratulations.

Vincent: Well, what about the interview?

Dr. Lamar: That was it.

Why interview when you know everything you want to know about someone?

They only needed authentication. Or can be the same today except no DNA check, just papers...

I recall a story of a suspect's dna that contaminated a crime scene by a paramedic. The paramedic had assisted the drunken suspect earlier in the day and transported him to a hospital. His hospital stay was his alibi.

There should be a way to expire the records from the database with exception perhaps for convicted felons.

Preferably not collect the data and not store it at all unless aforementioned.

I did a stint in federal prison. When I went to the halfway house, I had to wear an ankle monitor that reported my whereabouts 24/7. At first, I was a bit irked about it but in a few days there, I could see that it would be a benefit to me.

There were other guys at the halfway house who didn't have to wear one, and they were continually tempted to be places they shouldn't have. Too boot, they couldn't easily prove their innocence when they were falsely accused of being somewhere they weren't. I never once had that problem or temptation and cakewalked my stay there.

As long as the nation doesn't turn into a democratic socialist PRC, this technology will benefit people who are law abiding, even if a person doesn't have a record in some DNA database.

Unique identity markers aren't the property of an individual.

Can DNA be forged?

YES!!! A million times yes!

The way "DNA matches" occur in the judicial system is by looking at a set of SNPs, not whole genome comparisons (obviously).

This is how you frame someone.

1) Take the SNP set the cops use, which is published. 2) Make synthetic oligos that match the SNP set with SNPs from the person you want to frame. 3) Sprinkle the oligo collection at the crime scene 4) Guess what shows up in the PCR

s/SNP/STR/, and you're right. They're easy to manufacture.

Looks like they only look at 20 STRs typically. Also mtDNA and Y-chrom STRs. Maybe a little more challenging to where you would need to grow up some plasmids to toss around a crime scene, but nothing a first year grad student couldn't do in their garage for a few thousand bucks. Wouldn't even need to clean up the minipreps since it's PCR and you'll find exactly what you're looking for and nothing else :)

Not really, but it can be badly sampled or mismatched. Or misrecorded in databases.

Scott Greenfield, a practicing defense attorney, wrote about this and previous attempts at using "science" to create/identify "indisputable" evidence: https://blog.simplejustice.us/2019/01/22/junk-science-is-dea...

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact