Hacker News new | past | comments | ask | show | jobs | submit login
DNA Evidence Can Be Fabricated, Scientists Show (2009) (nytimes.com)
154 points by apo on Feb 2, 2019 | hide | past | favorite | 28 comments



DNA evidence combines two of the worst characteristics:

1) Taken as if it's the gospel of God to establish guilt

2) Very easily stolen/taken from a person and planted to a crime scene. Perhaps not for something like DNA found in a pool of blood on a fight scene, or semen at a rape, etc, but anything involving e.g. hair, nails, and other easily obtained body scraps...

As someone from an area of the world that used to have a large "deep state", police routinely planting stuff, and so on, I hate this kind of "proof".

People from Latin America, Africa, certain places in Asia, East Europe, etc will probably identify.

But even in the US, I've read tons of such stories, where the police had it for this or that person and planted evidence. Doesn't even have to be corrupt police, they just might be "certain" that "this n...a did it" without actually having evidence.


Can’t trust DNA, can’t trust videos or audio, certain state actors spewing constant misinformation...

We’re back to the 1800s in terms of the average person’s ability to know and/or prove the truth. Except this time around the state and some corporations have near-complete surveillance capabilities.


I actually wrote about that recently:

https://austingwalters.com/the-last-free-generation/

Basically, there is no freedom any more (if there was any to begin with). Without the ability to tell the truth, and with companies / governments having the ability to know & manipulate everything in your life, your life is not your own.


i'm receiving a 500 database error on your link FYI


We still rely on fingerprints which have never been proven to be unique, much less actual pseudoscience in the USA like "fire analysis." Might as well be dunking the accused to see if they are witches.


Don't forget about shoeprints, dental imprints, drug dogs, and other bunk


Hell, we actually executed a guy based on 'fire analysis.'


If you live under a system where the justice department cannot be trusted to find truth in good faith, then that is the problem. Not the tools at their disposal. It doesn't matter the tools. Every tool used to date can be abused.

But if DNA, video and audio can be used to help the justice department find truth, then I don't know why they get hated on so much. And if they are dishonest, I would rather make them go through the effort of faking DNA evidence and manipulating video instead of simply lying about confessions and coercing witnesses into making false statements.


Quantitative facilitation brings a qualitative change.

Even if you already have a justice department you don't trust, the abundance of tools that can be abused, can make the situation 10 times worse.


Maybe the state and corporations are developing surveillance capabilities because 3rd-party information can't be trusted.


Thanks for posting this. It’s surprising how few people are aware that this is possible.


It's more of interest for curiosity's sake than anything else. It's always been the case that a person could be framed with "DNA evidence" with the considerably more low tech approach of obtaining leftover hairs or bits of skin (or legally mandated DNA samples for testing) and depositing them at the crime scene, or simply finding a lab willing to lie about carrying out a test. DNA testing as evidence has always relied on the assumption the people collecting and testing the samples are honest (and competent)


Sorry for the long post but I felt it was related and contains personal experience in a criminal case tried by a jury that I took part in:

----------

This always comes to mind when I think of anything to do with "DNA Evidence" or "using Lab analysis or DNA sequencing" in terms of crime scene evidence as the "most important" or "most damning" portion / part of evidence brought against a defendant:

Annie Dookhan, chemist at Mass. crime lab, arrested for allegedly mishandling over 60,000 samples

----------

https://www.cbsnews.com/news/annie-dookhan-chemist-at-mass-c...

Like red light cameras, radar guns, drug dogs and anything else law enforcement and/or the government like to use to bolster their cases or work toward a more iron-clad / strong body of evidence to win their case it seems that DNA Evidence is something that was never really believable or strong beyond a shadow of a doubt.

I spent over a months time on a jury for a double attempted murder and while 95% of the testimony, expert analysis, and other direct / cross-examination was compelling enough to keep me focused and tuned in the lengthy, scientific, and overall information-packed expert testimony from the DNA Lab Scientist who performed the DNA Analysis did little to keep me awake.

Throwing endless numbers, percent chance that a sample is from a specific person, the likelihood that a sample could be from anyone beside a specific person, or the methods and means by which a scientist comes about determining the likelihood of a given sample belonging to a specific person seems like something [on paper] that should NEVER have been given enough credit to stand as evidence that could be believed beyond a shadow of a doubt. I felt certain, in those moments, that even the smartest folks on the jury and in the courtroom those days experienced little more than information overload, a lack of context, and certainly a sense of "I have to believe this because I'm not informed enough to question it and the world seems fairly certain that DNA Evidence is not only acceptable but highly accurate and therefore nearly useless to question".

It all felt and still feels like a grand performance akin to a magician using misdirection or sleight-of-hand to keep their audience paying attention to the exact wrong things.

----------

Lastly, I think it's also important to note that the defendant confessed immediately upon being picked up by the police without any lawyer present and it was all recorded on video and audio. The defendant was 99.999% guilty from all directions including analogical, anecdotal, character, circumstantial, demonstrative, digital, direct, exculpatory, forensic (beyond DNA), hearsay, physical, prima facie, and most damning: physical evidence.

I imagine in another case without such strong evidence beyond forensic evidence being the cornerstone of the State's evidence that the State would have liked to keep the case away from a Jury thus seeking a plea deal. I can only assume that would be the case, but from my POV in the ordeal the only reason the State spent so much specific time on forensic evidence was due to formality, due to due diligence, and because of the mountain of other credible evidence it served to bolster all of it.

Without any other hard evidence I can't imagine many juries taking only Forensic DNA Evidence as beyond a shadow of a doubt. No matter how certain or how close to 100% a scientist or lab analyst says a DNA Sample matches a certain persons the entire point of "without a shadow of a doubt" is to prevent innocent folks from winding up behind bars for things they didn't do when the people trying to get a conviction don't have the evidence, tactics, strategy or any other means to demonstrate the Defendant's guilt.


It's important people become aware of it since right now all the jury will hear is "there is a 1 in X octillion chance it is not this guy's DNA" and they will convict.


Many other things are possible, like breaking into a video surveillance system and replace the actual footage with a fake one incriminating someone else.


Much more difficult. You need to do a break in, and more importantly, to create a convincing fake footage.


It's crucial to remember that "DNA evidence" does not mean "DNA sequence". It's much less technical than that. Basically, you look at some set of sequences that vary a lot among populations and individuals. The choice of sequences, and the total number, varies. And it's generally increased over time.

So anyway, if they were doing a total DNA sequence, fakes would be a lot harder. Because you'd need to synthesize (or assemble from some library) the sequence for each chromosome. And that's still a lot harder than doing a complete sequence. For now, anyway.


>So anyway, if they were doing a total DNA sequence, fakes would be a lot harder.

Wouldn't they be as easy as taking some hair from a person and dumping it to the crime scene to be collected?

Doesn't even have to be the police, someone wanting to frame them could do it trivially.


True. But they'd need physical access to the target.

And actually, back in the day when I mailed cash, that was part of my OPSEC. I had a jar full of dust from public places. Working with disposable gloves, I'd put the cash in a new plastic bag, toss in some public dust, and shake well.


>True. But they'd need physical access to the target.

Or to just go (or have someone go) right after them at the hair saloon?


That's also physical access.

My point is that $TLA couldn't get DNA to implicate $TARGET in a murder, if they don't know where $TARGET is. Or know anyone who does. But if they can obtain $TARGET's DNA test results, they could fabricate evidence.


>That's also physical access.

We've already discussing the situation that someone is committing an actual crime for which they want to frame another. Compared to that, getting "physical access" to someone's hair saloon is a trivial part.

>My point is that $TLA couldn't get DNA to implicate $TARGET in a murder, if they don't know where $TARGET is.

Why wouldn't they know where $TARGET is? That's the easiest thing to know. That would only be a problem is they wanted to implicate fugitives already on the run or in hiding, or Waldo.


For alibi, in the future it would probably be a good idea to have a 24/7 personal body camera which timestamps the video using the blockchain or something.


A perfect way to counter misinformation from foreign and/or domestic hostile powers. Sign everyone up. Whoever refuses gets a hit to their social credit rating.


Would have to be self-contained; otherwise, the video could always be edited...


Obtaining the hash value of a video, then embedding that hash value into a secure block chain is one approach that gives evidence that a specific video (or any other document) existed before some point in time.

See, for example:

https://petertodd.org/2016/opentimestamps-announcement

Editing the video after the fact changes its hash value, thereby invalidating the timestamp on the altered video.


Perhaps camera sensors could digitally sign their output using an invisable watermark? It's been an idea I've had since deep fakes became a thing.


Doesn't really solve the problem, there are many ways of bypassing that, from having a small screen in front of the lens playing fake video, to hacking the firmware.




Join us for AI Startup School this June 16-17 in San Francisco!

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

Search: