The Reg and Carole McCartney tweeted about a New Jersey DNA case which I found very interesting because it intersected with a number of issues I’m working on at the moment. The case is about using DNA forensics to test a weapon that was allegedly thrown away by a murder suspect. A weapon was recovered and it was hoped DNA testing would prove that the suspect(s) had handled the weapon. The problem they encountered was that multiple Profiles (aka mixed profiles) were detected, further the other suspect in this case was not one of them.
Newspaper article on the case. Note: all suspects are considered innocent until convicted in a court of law.
There are two defendants in this case. Corey Picket had his DNA Profile detected but claimed that the gun the police found was not used by him and that he is innocent. The State of New Jersey used Technology called TrueAllele to determine whether he was a contributor to the mix of dna recovered. He challenged this result and demanded to inspect the code used in TrueAllele. Note: in legal cases it is always the defendant who is said to make the claim, even though it was probably his defense team. What was outstanding was that he won the case! Note the victory was from an Interlocutory hearing to an Appeal Court who remanded the case back to the lower court. The Appeal Court ruling is here [PDF]. The other accused was not declared a contributor to the Mix.
The USA is a Union
The US, the federal government entity, is said to have a system of Dual Sovereignty where Sovereignty is said to be shared between the US and the States. Each State is an independent Country and has it’s own legal system. The States have at least the standing of Scotland in the UK. This case is the first time that TrueAllele has been challenged in a New Jersey Court. There were a number of Innocence and freedom concerned organizations that were motivated to assist in the challenge. The Company behind TrueAllele (Cybergenetics) has invested a lot of money in creating their Software and has presented it as the best there is in untangling DNA mixes! They had an expectation that defendants would simply accept the result of their Software. Some defendants did challenge their result and asked to inspect the TrueAllele Code however Cybergenetics rejected that because their Code contained Trade Secrets. Note that other States are not obliged to accept the decision from New Jersey although they probably will.
I try to Blog covering the legal systems of the USA and GB. This is in some ways easier than you might expect because they both have a Common Law ancestry. Naturally there are differences as there are about 250 years that separate the two but in fundamentals they still retain a lot of similarity. In some ways the US holds on to aspects of the English legal system that have changed or been renamed in England. There is one relevant aspect that may be exactly the same in the US and UK however the standing and use of that aspect has changed. I refer to the Amicus Curiae. The name has Roman origins. In England they are presently called an Advocate to the Court.
The Amici will present a brief to the Court that concerns the current case where they have expertise. They are not Expert Witnesses, in the legal sense, or a Party to the case but usually legal scholars with an interest in that area. They usually submit their Brief in an Appeal Court. Interestingly the UK Courts also allow an Amicus Curiae (Advocate to the Court) Brief however there are differences [PDF]. In the UK they refer to the idea of Intervention. In the USA the Amici volunteer their Brief to the Court and the Judge may or may not accept it. In the UK an Intervener may be asked by the Court to help out in a specific area. Volunteering to be an Intervener is also known in the UK. The main difference is that this is very popular in US Courts but still quite rare in the UK.
The term Brief/brief is used in UK and US Law however it is not used in quite the same way. In the US a Brief is a document prepared in advance that is given to the Judge to help clarify the issues to be resolved. A Brief that is prepared by a non party to the case is called an Amicus Brief. Those preparing the Brief are said to be on-the-brief, which means their names are on the Brief. A Brief is not necessarily prepared by Lawyers but if the Judge allows it to be argued in Court an Attorney has to do that. This leaves a bit of wiggle room, is an Attorney hired to argue the Brief said to be on-the-brief? Confusing whether it means “on the document” or “in the team”. In the UK the term “Brief” usually applies to a Solicitor (possibly a Barrister) who advises a defendant. When a Barrister/Solicitor does not wish to be on a given Case anymore and they hand that over to another Lawyer they are said to handover the Brief. This confuses the issue of whether a Brief is a job or a role. In UK police stations when a suspect asks for a Brief, they know that person has experience in the CJS.
The Amici can be Lawyers or have Lawyers on their team. There were Lawyers from Harvard law School (Cyberlaw Clinic), Massachusetts Bar, Upturn and the Legal Aid Society (of New York), the Innocence Project plus a few others. Given that this is a New Jersey Case, it is quite exceptional to get so many out-of-State Attorneys admitted pro hac vice. It must have been regarded as a very important case. Normally an Attorney must be registered in the State (New Jersey in this case). A Judge can admit out-of-state Attorneys if they are in good standing with their registered States Bar.
It may see odd that such a powerful team can be assembled for such a horrible case. What I believe has happened is that the organizations in the Amicii were already working on the issue of gaining source code access to the Software and this case came along! The defendant/appellant must have thought it was his lucky day when interest was shown by such high powered legal Eagles!
This case has, in my humble opinion, indirect relevance to English Common Law: specifically a 1997 Law Commission recommendation that a common law presumption be used that a computer is operated correctly unless there is explicit evidence to the contrary. Unless you are familiar with this aspect of English Law you will be surprised that it could be true. It means that if Software (e.g. in an Accounting System or an Intoximeter) can be used to provide evidence against you in a Court, you can’t challenge it! This is discussed in depth in a book: Electronic Evidence by Stephen Mason and Daniel Seng. There is another good write up of the issues involved here [PDF]. In the USA each technology is judged on its own merits and sometimes code Access is granted. There is a presumption that Code is correct but it can be challenged. Recently in the Chun case access to source code was granted and found to contain bugs.
This Case may have more direct repercussions for the UK. Should TrueAllele be found to have bugs/errors that Software is used to help build the UK national DNA Database (NDNAD). It could cause many cases to be overturned but let’s not get ahead of ourselves as the Code Inspection is still some time in the future.
The State of New Jersey was yet to admit the type of DNA Testing offered by TrueAllele (Probabilistic Genotyping). It was accepted that a Frye hearing was required to do this (Frye v. United States, 293 F. 1013 (D.C. Cir. 1923)). The objective of the defendant was to use the Frye hearing to reject this technology in New Jersey. The State put forward an expert to determine the technology was reliable and further that no Code Inspection was required, that the result should just be accepted. The expert also stated that the Code was a Trade Secret and the Company would suffer adverse financial consequences if those secrets were ever released.
Summarizing
The amicii in this case had a summary of their Briefs written up, I will attempt an even shorter summary
(i) New Jersey Attorney General (AG). The AG argues that access to the source Code is not required. Their expert is Dr. Perlin who is a founder of Cybergenetics.
(ii) The Innocence Project (Inc.)
Their Amicus Brief underscores these indisputable facts: people write source codes; people make
mistakes.
The Innocence Project states that genotyping software is prone to error,
as exemplified by the problems associated with STRmix and FST. Note: so many problems were found in FST that is was abandoned.
The integrity of STRmix has been challenged in the UK, specifically in this case [PDF] and [PDF].
(iii) Upturn.
An organization seeking to advance equity and justice in the design, governance, and use of technology.
TrueAllele’s source code has never been independently reviewed, and that such a review is a basic and necessary step in ensuring reliability
(iv) The Association of Criminal Defense Lawyers of New Jersey (ACDL-NJ)
An established New Jersey Firm, asserts that probabilistic genotyping has not yet been used or
tested in New Jersey. TrueAllele’s leading competitor, STRmix, has produced its source code and
conceded its software had errors, rigorous scrutiny of TrueAllele’s source code becomes even more compelling.
(v) The Legal Aid Society
LAS is the primary public defender in New York City.
LAS has firsthand experience litigating the admissibility of a proprietary probabilistic genotyping program—FST—including successfully obtaining access to FST’s source code.
FST was scrapped.
(vi) Mats Heimdahl and Jeanna Matthews
Drs. Heimdahl and Matthews are experts in engineering, testing, and validating computer systems, including forensic evidentiary software.
(vii) American Civil Liberties Union of New Jersey
The ACLU-NJ argues that independent review is essential. Questioning Dr. Perlin, reviewing validation studies and peer-reviewed articles in which he or his current or former employees were involved, or relying on out-of-state judicial opinions citing his testimony and those studies misses the importance of objective analysis of the science underlying his forensic testimony.
ACLU-NJ raised the issue of a Likelihood Function. Most non-experts in forensic DNA technology believe that it gives you a binary result. It either is somebody’s DNA or it isn’t. In practice it is very difficult to obtain a result as clear cut as this. Instead they need to use probability and statistics. Even if DNA is collected and processed according to guidelines it remains a challenge. The standard technique is to measure some number of predefined Alleles, particularly those that are classified as Short Tandem Repeats (STR). The frequency of these in the community is used, with the probability multiplication rule, to produce a very low number that is then inverted to estimate a probability. The numbers generated can be huge, they may imply that a number of additional planet earths are required before someone with that same profile is encountered again. Sensing this is nonsense the UK regulators put a cap on the odds to 1 billion to one. In the US people come up with numbers of multiple quadrillion to one. In practice the measurement of Alleles is not perfect, there are drop-outs, drop-ins and blobs! This means people try to come up with an estimate using that imperfect data. Getting back to the current case, not only will you have those problems you may have multiple contributors to that sample! This is known as a Mix and is the sort of problem that TrueAllele attempts to solve. A result of this is that it has to produce a Likelihood Ratio. This is not even a probability that someone is a match.
ACLU-NJ raised this point in their Brief but they are only an amicus and as none of the Parties raised this point the Court decided to not discuss it. It is still left to the Parties to raise it later.
Discussion
There has been a lot of discussion about obtaining the source code in multiple cases. These tend to focus on the correctness (or otherwise) of the code. More recently there has been a shift of focus to the programmer. One reason is that programs can often comprise many thousands of lines of code. The State’s expert witness in this case estimated it would take about 8 years to code review their software. Another approach is to review the Programmers! In Martin Fowler’s book: Refactoring he introduces the concept known as “Code smells”. The basic premise is that code should be professionally written using a set of highly regarded techniques. His book was very popular around the turn of the century, less so now with the widespread use of JavaScript packages (E.g. Angular, React). The idea is that you would not need to review all the lines of Code to determine that was badly written. It is also popular in some development frameworks to do automatic checking of the Rules as you develop.
TrueAllele is estimated to contain about 170,000 lines written in MATLAB. MATLAB is not generally used write software, it’s main focus area is electronic systems. In a University MATLAB would be popular in the Electronics department whereas Java would be more popular in the Computer Science department.
In US cases an important consideration are your Rights. The State of New Jersey will have a Bill of Rights (and Constitution). The US has a Constitution (with embedded Bill of Rights), it is also possible some historical vestiges from the English Bill of Rights survive. There will also be a lot of decisions from the respective Courts (Case law). In the UK you also have Rights but they are spread over many existing Laws, historical and modern, and it can be difficult to determine what they are. e.g. you have a Right of Silence but the prosecution can infer that you have something to hide from your silence. So what sort of Right is that?
The right to a fair trial is fundamental and guaranteed pursuant to the Fifth and Sixth Amendments of the United States Constitution, as well as the New Jersey Constitution
This implies suspects have the Right to challenge the evidence against them.
The following link provides an excellent survey of the issues associated with a Constitution and Software.
CRIMINAL DEFENDANTS’ CONSTITUTIONAL RIGHT TO SOURCE CODE (PDF)
The most interesting point is your right to confront the witnesses against you. The question is who is a witness in the modern world? We have historically presumed that they are people however we now have a lot of software that is used as evidence against us. How do you challenge a programmer who made an error years ago perhaps in another Country? The authors make the point we should reconsider who or what is a witness and be allowed to challenge by having a code an adversarial code inspection.