Michigan’s evidentiary breath alcohol testing system has undergone a significant and historically uncommon transition. The Intoxilyzer 9000, manufactured by CMI, Inc.,¹ is now the evidentiary breath testing instrument in use across Michigan, replacing the DataMaster DMT and concluding a lineage of breath testing technology that has been in continuous service in Michigan since the 1980s.² While updates to breath testing instruments are familiar to most practitioners, the adoption of the Intoxilyzer 9000 represents a departure of unusual institutional and legal significance.

The replacement of the DataMaster platform with the Intoxilyzer 9000 marks the first time in nearly half a century that Michigan has adopted an evidentiary breath testing instrument produced by a different manufacturer.³ The result is not merely new hardware but the introduction of a structurally different yet strikingly similar technological architecture.

Any transition of this magnitude carries practical and jurisprudential implications for courts and practitioners. With the adoption of a new and different evidentiary platform, Michigan’s Administrative Rules had to be updated; the DMT caselaw is now subject to renewed examination, and longstanding doctrines may require reconsideration or refinement.

This transitional period represents a unique moment for the Michigan legal system, one in which courts, prosecutors, and defense counsel alike will play a role in determining how existing precedent applies to the new technology and whether portions of that jurisprudence will endure, adapt, or give way to revised legal frameworks governing the admission and evaluation of breath test evidence in OWI litigation going forward.

INTRODUCING HISTOGRAMS: ENHANCING TRANSPARENCY WHILE CHALLENGES PERSIST

The Intoxilyzer 9000 (I-9000) promises greater subject sample transparency largely due to its ability to generate a histogram of the subject’s breath testing sequence. The histogram visually represents the subject’s breath flow rate on one vertical axis and the breath alcohol content (BrAC) on the other, with elapsed time shown on the horizontal axis. Ideally, the histogram provides practitioners with an additional tool to assess the validity of the test results and identify potential issues.

In practice, however, defense attorneys must scrutinize these histograms for inconsistencies or anomalies. This is often difficult due to the small size and poor clarity of the printed histograms, which apparently cannot be provided on a larger scale. A proper histogram typically shows the subject’s breath flow rate increasing, leveling off, and then decreasing.⁴ Deviations from this expected pattern, such as a failure to level off, may indicate the presence of mouth alcohol.

Although histograms are generally produced for each test attempt, defense attorneys have observed exceptions where histograms of all sample delivery attempts are not provided. There is an important distinction between the device generating a histogram and the defense receiving access to it. In some cases, attorneys report not receiving histograms for all attempts, raising questions about whether the missing data could reveal issues with the testing process. When additional histograms are believed to exist, they can be requested and individually printed directly from the I-9000 itself, as the full test sequence is stored in the instrument’s internal memory. This process requires cooperation from law enforcement, as these histograms will not be automatically collected through the MSP’s “COBRA” data, which is discussed below.

In other cases, the histogram appears to start in the middle of several attempts, so the breath flow does not start from zero on the y-axis. This makes it difficult to determine whether there is a spike in breath flow or an indication of mouth alcohol. When MSP was asked whether the I-9000 had the ability to go back and show the entire attempt, it said it could not do so.⁵

Further complicating matters, histograms do not always reflect the expected patterns. For instance, there are cases where the histogram suggests the presence of mouth alcohol, as evidenced by a BrAC peak followed by a rapid decrease, yet an ethanol reading is still reported.⁶ Conversely, there are also instances where the histogram appears consistent with an acceptable sample, but the breath test ticket indicates that the sample was rejected with the status code “unacceptable sample.” This is unsurprising, as the MSP training materials and the Operator Guide⁷ acknowledge that this “exception message” can result from several potential causes, including residual mouth alcohol, yet it cannot definitively confirm or refute its presence.

These discrepancies highlight the importance of thoroughly examining the histograms to determine whether external factors, such as residual mouth alcohol, environmental contaminants, or internal errors, such as calibration or operational issues, could explain the inconsistencies. Ultimately, while the I-9000’s introduction of histograms marks a step toward greater transparency, defense attorneys must remain vigilant. Gathering all the histograms and carefully reviewing them, corroborating evidence such as video recordings of the breath test sequence, and the existence of potential outside influences is essential to ensure accurate and fair outcomes for defendants.

DIFFERENCES IN PROVIDING ACCESS TO DATA

In addition to the newly included histogram, the I-9000 BrAC printout will provide more information than was available on the DMT printout. The I-9000 is intended to provide easier access to this information, as the MSP have purchased Computer Online Breath Alcohol (COBRA) software, a data-management system designed to collect, archive, and retrieve detailed electronic records generated by Intoxilyzer breath-testing instruments. While the MSP plans to link all the machines into a local area network so that it can obtain the data from all machines, this has not happened in the two plus years since the I-9000’s rollout. Other states like Texas,⁸ Arizona,⁹ and Georgia have not only linked the machines in a network but have provided defense attorneys with the ability to directly access the data from any machine online. For now, Michigan practitioners will have to request the data through FOIA.

DIFFERENCES IN ACCURACY CHECK PROTOCOLS

The I-9000 represents a significant departure from the accuracy verification framework used with the DataMaster DMT. Under the DMT system, evidentiary reliability was supported by multiple layers of periodic testing, including weekly dry-gas accuracy checks and quarterly wet-bath calibration checks. Those protocols have now been replaced.

Rather than relying on separate weekly or quarterly calibration procedures, the Intoxilyzer 9000 verifies calibration by running a dry-gas standard check as part of each subject test sequence. The I-9000 will continue using the same dry-gas standard as the DMT, sourced from Guth Laboratories with a nominal target rate of 0.080.¹⁰ However, the acceptable range for a dry-gas test to be deemed acceptable is increasing from 0.076-0.084 (+/- 0.04 either way) for the DMT to 0.074-0.084 (+/- 0.06 below and +/- 0.04 above) for the I-9000.¹¹ No studies have been presented, and no explanation has been given for the increase in the acceptable range for the same dry-gas standard that was being used with the DMT or for why the total acceptable range is now an asymmetrical 12.5% rather than 10%, with 5% each way.

Because the Intoxilyzer 9000 no longer undergoes separate weekly accuracy checks, review of an instrument’s performance necessarily depends on examination of dry-gas checks embedded within subject test sequences. In practice, this means also evaluating the dry-gas results associated with tests conducted before and after the defendant’s test on the same instrument, which are obtainable through FOIA.

Another accuracy-related change to the rules governing breath tests is that the I-9000 will not have to undergo the 120-day maintenance and accuracy testing to which the DMT was subjected. Instead, the machines will be required to undergo the maintenance and accuracy testing “not less than two times annually.”¹² The I-9000 printout will list the date of its most recent 182-day test, which should be obtained by defense attorneys via FOIA request to the MSP. One would assume that those records will be a required foundation to the admission of the I-9000 test results, but that remains to be determined by the courts at Daubert¹³ hearings.

There is nothing in the new rules stating what type of tests must be run during these 182-day inspections. Rather, the rules require only that the machines “be inspected, verified for accuracy, and certified as to their proper working order.” However, in December 2025, the MSP finally issued the protocol¹⁴ for the biannual inspections, and it is similar to the prior protocol for the DMT.

Regardless, the new rules require only that a dry-gas sample with a target of 0.080 be run with each subject test and that the machines be inspected twice per year.¹⁵ There is no requirement for testing at targets above 0.080, such as the 0.17 threshold for OWI–High BAC, but the biannual inspections require testing at 0.040, 0.080, and 0.200. Still, since the I-9000 is testing only a 0.080 within each subject test, questions remain about how the MSP will ensure the accuracy of tests above 0.080 in between the two annual inspections. The diminished testing requirements suggest that the MSP has placed unwarranted confidence in the I-9000’s ability to maintain accuracy without additional oversight, a concerning assumption given the potential implications for high-BAC cases and overall reliability.

DIFFERENCES IN OFFICER TRAINING AND CERTIFICATION

All law enforcement officers operating the I-9000 originally completed training and certification on the device. The training process included a two-step program: Officers first reviewed an online PowerPoint presentation and passed an online test to demonstrate their understanding, and then they completed hands-on demonstrations on an actual I-9000 machine.¹⁶ Michigan has also published an official Intoxilyzer 9000 Operator Guide, providing officers with detailed instructions on the use and interpretation of the device, including guidance on breath graph analysis, exception messages, and operational procedures.¹⁷

Another change from the DMT is that officers are required to be recertified on the I-9000 every two years, and the officer’s certification date appears on the I-9000 ticket.¹⁸ The recertification process requires the officer to view a slide show and take a written test, but it does not require the officer to demonstrate proficiency with a hands-on test. If the officer’s certification has lapsed, the I-9000 is not supposed to allow the officer to proceed with the test. However, the officer could manually enter a different certification date. It has been more than two years since the I-9000 was rolled out, so Defense counsel should either obtain a copy of the officer’s certification date independently from the I-9000 ticket or require the officer to produce his or her training certificate at a hearing or trial to verify that the officer did not incorrectly enter the date of his or her training.

NO DIFFERENCES IN DEPRAVATION-OBSERVATION PROTOCOL

Something that did not change with the I-9000 is one of the most litigated areas concerning breath tests: the 15-minute observation-deprivation period.¹⁹ The I-9000 has the capability to require the officer to start the testing sequence and then lock them out for 15 minutes to satisfy the observation-depravation requirement. However, the MSP chose to disable this feature, allowing officers to begin the 15-minute observation period before the I-9000’s clock starts tracking, thereby avoiding an additional 15-minute wait when another officer and subject are still completing their testing.²⁰ This decision was made despite the MSP’s stated “best practices,” which recommend that officers rely on the I-9000’s clock to determine the start of the observation period.²¹

DIFFERENCES IN CALCULATING AND DISCLOSING MEASUREMENT UNCERTAINTY

Unlike the DMT, where the MSP calculated only an “average” uncertainty budget using just a few machines, the measurement uncertainty of each I-9000 machine is calculated after being placed in the field.²² However, that uncertainty budget is not included on the breath test printout and should be obtained from MSP or the local law enforcement agency. It is uncertain what factors go into the measurement uncertainty calculation, whether the individual making that calculation is qualified to do so, and whether the calculation includes both Type A uncertainty and Type B uncertainty.²³ In addition, it is unclear how often the uncertainty budget is updated after the initial calculation and who is responsible for ensuring that the update occurs. All of this is fertile grounds for cross examination at a Daubert²⁴ hearing on the admissibility of the I-9000’s test results.

COMPREHENSIVE DISCOVERY STRATEGIES FOR THE INTOXILYZER 9000

The transition to the Intoxilyzer 9000 has introduced new challenges and opportunities for defense attorneys handling DUI cases in Michigan. The Michigan Forensic Breath Alcohol Analytical Report (MFBAAR), which replaces the OD-80 used with the DataMaster DMT, is now the foundational document provided at the conclusion of a breath test. This report includes the subject’s Breath Alcohol Content (BrAC), duration of the blows, dry-gas calibration accuracy test, number of attempts, and total breath volume. For practitioners, obtaining and carefully examining this document is the first step in discovery.

The MFBAAR alone rarely tells the entire story. Each Intoxilyzer 9000 machine stores a wealth of additional data within its internal memory, much of which may someday be accessible through the COBRA software. This data includes calibration checks, exception messages, and historical test results. While direct access to COBRA data is not yet available online in Michigan, it is an invaluable resource for identifying patterns or irregularities in the machine’s operation. Defense attorneys should routinely request this information through both discovery and FOIA submissions to the MSP. Specificity in these requests is crucial; attorneys must include details such as the machine’s serial number, the MFBAAR report number, location of the instrument, operator’s name and certificate number, the date given, and the location of the test.

Histograms provide a valuable tool for correlating visual data with the numerical results in the MFBAAR. As previously discussed, when additional histograms are believed to exist, such as for aborted or rejected test attempts, they can be requested and individually printed from the I-9000’s internal memory. Ensuring access to this data is critical for constructing a comprehensive defense.

Attorneys should request calibration data for the specific test sequence in question, including the dry gas standard’s certification of analysis. This certification ensures the accuracy and reliability of the gas used in calibration. The absence of a valid certification may undermine the foundation for admissibility of the test results and should be carefully examined. Attorneys should also request records of historical subject tests from before and after their clients’ tests to review the dry gas calibration checks on those tests as well.

Exception messages, printed on the MFBAAR, document conditions under which the Intoxilyzer 9000 determines that a subject sample, instrument function, or testing environment falls outside acceptable parameters. As defined in the MSP Operator Guide, these include, among others, deficient or unacceptable samples that fail minimum requirements for pressure, duration, volume, or slope; purge failures and ambient failures indicating contamination of room air; calibration checks out of tolerance; detection of interferents or radio frequency interference; and diagnostic failures reflecting internal system checks that did not pass. These messages do not identify the root cause, but they record that the instrument detected a deviation sufficient to interrupt, reject, or qualify the testing sequence.²⁵

Historical exception-message data for the specific instrument, obtainable through FOIA, must therefore be carefully evaluated in context. Attorneys should examine whether an instrument exhibits an unusually high number of multiple attempts, repeated deficient or unacceptable samples, deviations from typical breath volumes, frequent purge or ambient failures, or repeated detections of radio frequency interference or other interferents. Patterns of such events, particularly when clustered over time, may indicate environmental conditions, overly restrictive or unstable sample-acceptance parameters, or declining instrument performance rather than isolated subject behavior.²⁶

Critically, exception-message patterns should be correlated with maintenance and service records. Where repeated exception messages appear without corresponding service calls, diagnostic intervention, or corrective maintenance, the data may suggest that known performance issues were not addressed in a timely manner. Under the MSP framework, service is initiated only when reported, and the absence of service activity despite recurring exception conditions may bear directly on reliability, foundation, and admissibility.²⁷

Finally, maintenance and certification records for the I-9000 must not be overlooked. As previously noted, the I-9000 follows a biannual certification schedule rather than the 120-day inspection requirement of the DMT. The dates and results of these inspections are foundational to the admissibility of test results and should be requested as part of any discovery demand. Additionally, records of software updates or repairs performed on the device may reveal potential sources of error that could impact the reliability of test results. In the absence of direct online access to COBRA data, discovery best practices for the I-9000 require thoroughness and persistence.

Accordingly, until centralized COBRA access is implemented, effective defense practice in I-9000 cases demands a deliberate and methodical discovery approach, one that combines persistent use of FOIA, targeted prosecutorial demands, and rigorous examination of MFBAARs, histograms, video recordings, and all associated maintenance, certification, and software records to ensure that breath test evidence admitted in OWI litigation is both scientifically reliable and legally sound.