Alex McCrory writes of the PSNI playing the same dirty tricks that became such a defining feature of it  back in the day when it went under its old name the RUC.

Recently an article appeared on The Carnary about my ongoing trial written by Peter Kearney. It covered some of the evidential aspects of the case relating to voice analysis and the destruction of evidence. I feel some further elaboration of the issues is required as the trial reaches a critical stage. It is only natural that I come at this from a defence perspective.

There were two voice experts for the Crown in this case: Professor French and Dr Kirchubble. How two experts came to be involved was as a result of Professor French misidentifying a voce of the tape, allegedly that of Harry Fitzsimons, with that of another person not before the court. Dr Kirchubble was brought in to explain how this fundamental error occurred. As she was an employee of French at the time, she was basically attempting to correct her boss's mistake. Unsurprisingly, she reached the same conclusions saying the mistake was with an 'margin of error'. A convenient excuse, if I may do so.
 
But the inherent flaws of the voice analysis is deeper than the misidentification. From the outset, the experts were provided with a police transcript that named the three defendants as suspects. Additionally, large portions of speech was attributed to each of the three accused, and all of this before the experts commenced their analysis. In their evidence, the experts admitted to having had relied of the police transcript throughout the course of their work, with the exception of the first of seven drafts that were produced. 

As the analysis was being carried out, French and Kirchubble partook in a 'collective listening' exercise with the investigating officers. The evidence showed they had discussed the recordings and compared notes. Neither the expert nor police officers kept a record of precisely what occurred at that meeting. But, clearly, they were in cahoots.

At the end of giving his evidence, Prof French opined that he would not do the voice analysis the same way today. When asked by the judge why not, he said he would be careful to remove contextual bias, i.e. he would not introduce the police transcript until later stage in the process. Kirchubble also accept that she was susceptible to the same cognitive biss because of her use of the police transcript when compiling her reports. 

The role of an expert is to provide independent and impartial evidence to the court free from any external influence.

Another point touched on in the article is the defence claim that MI5 destroyed the original evidence. As with the Craigavon Two, the devices in this case were wiped clean within 24-hours of being retrieved and downloaded onto USB sticks. The net effect of this is that it is impossible to compare what was originally captured on the devices with the contents of the USBs. 

Strangely, the MI5 operative who downloaded the devices had no personal recollection of wiping them, although he said it had to have been wiped by him in the circumstances. According to his evidence, it is a matter of MI5 policy to wipe devices for redeployment purposes. Apparently, MI5 is not concerned with the retention of evidence which is a basic legal requirement. 

And it gets better: the computer that was used to download the recordings onto the USB sticks was also destroyed. Short of having access to the devices themselves, which the court previously ruled out, the computer was the only other link between the devices and the USB sticks. All this has put the defence in the impossible position of not being able to independently verify the recordings. 

The Canary article mentions the discovery of malware/viruses on the two of the USB sticks, but not on the third one, which has simply vanished into thin air. Since all three USBs allegedly were created at the same time using the same computer, the absence of the viruses on the third stick is a mystery. It is suggestive of a separate process about which nothing is known. 

Basically, the malware offered third-party access to the computer and/or the devices with the ability to edit or alter the data. A defence expert in the field of computer technology opined that the very existence of the viruses undermined the integrity of the evidence. 

These are some of the matters being disputed; there are others.

Alec McCrory is on trial in a Diplock Court
in Belfast where PSNI evidence tampering is the real story 

PSNI & Evidence Manipulation

Alex McCrory writes of the PSNI playing the same dirty tricks that became such a defining feature of it  back in the day when it went under its old name the RUC.

Recently an article appeared on The Carnary about my ongoing trial written by Peter Kearney. It covered some of the evidential aspects of the case relating to voice analysis and the destruction of evidence. I feel some further elaboration of the issues is required as the trial reaches a critical stage. It is only natural that I come at this from a defence perspective.

There were two voice experts for the Crown in this case: Professor French and Dr Kirchubble. How two experts came to be involved was as a result of Professor French misidentifying a voce of the tape, allegedly that of Harry Fitzsimons, with that of another person not before the court. Dr Kirchubble was brought in to explain how this fundamental error occurred. As she was an employee of French at the time, she was basically attempting to correct her boss's mistake. Unsurprisingly, she reached the same conclusions saying the mistake was with an 'margin of error'. A convenient excuse, if I may do so.
 
But the inherent flaws of the voice analysis is deeper than the misidentification. From the outset, the experts were provided with a police transcript that named the three defendants as suspects. Additionally, large portions of speech was attributed to each of the three accused, and all of this before the experts commenced their analysis. In their evidence, the experts admitted to having had relied of the police transcript throughout the course of their work, with the exception of the first of seven drafts that were produced. 

As the analysis was being carried out, French and Kirchubble partook in a 'collective listening' exercise with the investigating officers. The evidence showed they had discussed the recordings and compared notes. Neither the expert nor police officers kept a record of precisely what occurred at that meeting. But, clearly, they were in cahoots.

At the end of giving his evidence, Prof French opined that he would not do the voice analysis the same way today. When asked by the judge why not, he said he would be careful to remove contextual bias, i.e. he would not introduce the police transcript until later stage in the process. Kirchubble also accept that she was susceptible to the same cognitive biss because of her use of the police transcript when compiling her reports. 

The role of an expert is to provide independent and impartial evidence to the court free from any external influence.

Another point touched on in the article is the defence claim that MI5 destroyed the original evidence. As with the Craigavon Two, the devices in this case were wiped clean within 24-hours of being retrieved and downloaded onto USB sticks. The net effect of this is that it is impossible to compare what was originally captured on the devices with the contents of the USBs. 

Strangely, the MI5 operative who downloaded the devices had no personal recollection of wiping them, although he said it had to have been wiped by him in the circumstances. According to his evidence, it is a matter of MI5 policy to wipe devices for redeployment purposes. Apparently, MI5 is not concerned with the retention of evidence which is a basic legal requirement. 

And it gets better: the computer that was used to download the recordings onto the USB sticks was also destroyed. Short of having access to the devices themselves, which the court previously ruled out, the computer was the only other link between the devices and the USB sticks. All this has put the defence in the impossible position of not being able to independently verify the recordings. 

The Canary article mentions the discovery of malware/viruses on the two of the USB sticks, but not on the third one, which has simply vanished into thin air. Since all three USBs allegedly were created at the same time using the same computer, the absence of the viruses on the third stick is a mystery. It is suggestive of a separate process about which nothing is known. 

Basically, the malware offered third-party access to the computer and/or the devices with the ability to edit or alter the data. A defence expert in the field of computer technology opined that the very existence of the viruses undermined the integrity of the evidence. 

These are some of the matters being disputed; there are others.

Alec McCrory is on trial in a Diplock Court
in Belfast where PSNI evidence tampering is the real story 

16 comments:

  1. Sending best wishes from Australia

    ReplyDelete
  2. And the video the spooks have?

    ReplyDelete
    Replies
    1. One they made themselves? Would you believe a spook any quicker than you would Dodgy Danny Morrison?

      Delete
    2. That's a fair point. But I'd still ask why the spooks are so intent on fitting the three of them up though. Seems an awful lot of effort. And is Morrison is still on the Bobby Sands Trust board? Surely that's just offensive to the Sands family.

      Delete
    3. Probably because the spooks believe they are key players. But the spook rule of fitting them up seems to me to be the key issue. Those who insist on the rule of law can't then support the rule of law enforcement.
      Now, Dodgy Danny is not going to care in the slightest what the Sands family think. His withholding of information to the hunger strikers was a key factor in their deaths. If only he had been so eager to withhold information from the Brits.

      Delete
    4. Well it's obscene what happened to the hunger strikers if he really did allow them to die unnecessarily. I may be opposed but I do respect them for their courage. Something Morrison clearly lacks along with integrity.

      In regards to the OP, have any of the three denounced the use of violence or at least displayed an understanding of it's futility?

      Delete
  3. Neither Expert may have been biased but they do flag that their decisions were prone to bias -the rule of thumb is that the 'appearance of bias' is enough to undermine reliability; see Pinochet No. 2 judgment, which has been followed in scores of other judgments.

    And, 4 out of 10 miscarriages of justice involve unreliable witness testimony, see: https://evidencebasedjustice.exeter.ac.uk/current-research-data/testimony-evaluation/

    ReplyDelete
    Replies
    1. Christy - do you have a link to that Pinochet judgement?

      Delete
    2. Ordinarily arguments about the appearance of bias relate to judges but the reason I think it is relevant here is because voice recognition is not scientifically objective evidence but involves significant human judgment or lay-listener evidence. Both lay-listeners knew more that they were supposed to that it could have prejudiced their minds and the gap between when the recordings were made and they heard them could also be relevant -along with fact that originals have been destroyed which could allow for some editing or frequency manipulation that can alter the sound of voices. I think the natural background environmental static that an original recording would pickup can then reveal tell-tail signs of editing -and that is lost in a copy version because it becomes masked by the recording instrument static.

      Voice recognition basically falls into voodoo science like bite mark recognition which has been discredited.

      Here are links to Pinochet and another more recent judgment as well as an artical on voodoo science.
      https://www.bailii.org/cgi-bin/format.cgi?doc=/uk/cases/UKHL/1999/52.html&query=(pinochet)
      https://www.bailii.org/ew/cases/EWHC/QB/2019/2988.html
      https://www.wsj.com/articles/rejecting-voodoo-science-in-the-courtroom-1474328199

      Delete
  4. It not that the experts are biased per se. Bias was introduced into the process by the police transcription and other contextual information. However, the experts freely partook in a two hours "collective listening" exercise along with the prosecution authorities. This patently undermines their independence and impartiality.

    Alex

    ReplyDelete
  5. The role of the expert is provide independent and impartial evidence to the court. Although commissioned by the investigating agency to conduct analysis, the duty is to the court.

    Alex

    ReplyDelete
  6. Alex

    What you are suggesting is that they may have been improperly prejudiced rather than starting out with bias? If they listened to the tapes in the precence of the prosecution or cops then that would be highly irregualr and improper -because you could not rule out coaching or confirmation bias because you wouldn't know if someone was talking them through the recording -for example saying things like 'this next voice is McCrory', etc.

    ReplyDelete
  7. From the outset the experts were provided with a heavily annotated police transcript which they relied upon for 6 out of 7 drafts. The "collective listening" lasted two hours and, while there is no contemporaneous record of what took place, we do know they compared notes. All this came out during cross-examination. The experts were also supplied with contextual information which they simply do not require in order to conduct their analysis. Towards the end of his evidence the key expert said he would do the report differently today, being careful to remove the potential cognitive bais. This was a major concession by any stretch.

    I have explained all this in the original post.

    Alex

    ReplyDelete
    Replies
    1. Cognitive/contextual bias introduced by a written transcript is not critical to their evidence of actual voice recognition. I imagine the court will be more impressed with their self-effacing candour on a side issue rather than as you seem to be interpreting it.

      Delete
  8. Clearly you are not aware of case law in this part of the world. I refer you Flynn vs St John, a case in which cognitive bais arise from an annotated police transcript played a major role. The Court Of Appeal in thst instance overturned a conviction because police witnesses share an annotated transcript which, in the Court's view, give rise to the possibility of cognitive bais. Do, what you describe as a side issue was central to this ruling.

    Alex

    ReplyDelete
  9. I hope you are right but I would not hold out much hope that the court will treat Flynn v St John on a par with your own case. In Flynn the Voice Analysit concluded that the police recordings were not suitable for speech analysis and thus the police transcript was intended to make up for the deficit -thus the bias would have been more problemmatic than in your own case.

    Whereas, in your case both Analysists have indicated that the quality of the recordings were moderate to very strong. What this means for your case is that neither Analysist needed to rely upon the transcript because the audio was good enough for them to listen to for themselves and compare it with the samples. I think the court will see it as I have suggested but I could be wrong.

    ReplyDelete