Being Intolerant Depends on Humans not the Machine
Purpose: To challenge the Crown's assertion that the AI will automatically shut down if anything is out of tolerance or ATC or manufacturer's specifications.
Let's suppose the Crown's expert says this about the Intoxilyzer 8000C:
"If the calibration check is outside the acceptable range, you’re not going to be able to proceed with testing."
The expert is saying in other words, the instrument is smart enough to make sure that cal. checks are foolproof, and you can't proceed with testing unless the cal. checks are within the acceptable range of 90 to 110. He or she is saying there is a failsafe mechanism and Judges will interpret the evidence that way.
Two big problems (at least):
1. Do you have good 100 (or 80) standard (wet bath or dry gas) that has not been contaminated, for which continuity can be proven by proper documentation, that matches the Certificate of the Analyst, at 34C if wet bath and at the proper atmospheric pressure if dry gas, with proper tubing, and
2. On Ontario's 8000Cs did the previous breath tech who changed the bottle do his or her job properly when they MANUALLY SET THE TOLERANCE?
See Training Aid page 86 of 238:
But what if they enter 12, 15, or 18?
THE TOLERANCE IS SET BY A LOWLY QUALIFIED TECHNICIAN, not by the hardware, the software, the factory, or the Canadian Authorized Service Centre.
Q. Right. But the only reason why it is 10 –
you said that the instrument won’t allow it to proceed with
testing. The only reason why the limit is 10 in terms of the
instrument proceeding with testing is because the last
operator who keyed in information for purposes of change of
simulator solution keyed in the number 10. If the officer
keyed in the number 15 or the number 20 then it would be
completely different. It’s not the instrument that makes the
determination. It’s the operator who changed the solution
that makes that determination of tolerance.
A. The tolerance is set. It can be changed at
any time, but it’s supposed to be set to 10 when the alcohol
standard solution is configured.
Q. Okay, so it’s supposed to be set at 10.
Q. But if officers are not setting it at 10...
A. They have no reason to change it. They’ve
always been told it’s 10, that they shouldn’t be changed.
Q. All right. So, they’re training is it
should be 10.
Q. This is not a function of the instrument.
It’s a function of humans.
A. Well, the information is already
prepopulated into the questions.
A. And so, they just hit ‘enter’. There’s no
reason to change that result.
The cross-examiner needed to take this issue further. If there is "operator error" by a previous qualified technician who changed the solution (who isn't likely to be a witness at trial) then the Crown's hypothesis of failsafe instrument shut down at cal. checks outside of 90 to 110 can be disproven empirically. To disprove empirically you would need defence access to an 8000C with Ontario software.
The problem is potential for human error - it is not a failsafe system.
See also the CalGuard system in the list of tags at this site.