People v. Ambrosio: Counsel Not Ineffective for Failing to Seek an Unsettled, More Demanding “Drug Impairment” Jury Charge

1. Introduction

People v Ambrosio arose from a traffic stop in which the arresting officer reported indicia of drug use (odor of “fresh marijuana,” constricted pupils, droopy eyelids, argumentative demeanor), and recovered a THC vaping cartridge and “multiple loose suboxone medication strips.” Defendant Jason J. Ambrosio admitted taking suboxone hours earlier but denied marijuana use that day; he declined a drug influence evaluation and no quantitative toxicology evidence established the amount of drugs in his system at the time of driving.

A jury convicted him of (1) driving while ability impaired by a drug (Vehicle and Traffic Law § 1192 [4]) and (2) driving while ability impaired by the combined influence of drugs (id. § 1192 [4-a]).

The appellate dispute did not primarily challenge the sufficiency of proof; it targeted the jury instruction. Specifically, defendant claimed ineffective assistance because trial counsel requested the then-“standard” model instructions (tracking People v Cruz’s “impaired to any extent” formulation) rather than asking for a “heightened” drug-impairment definition described by the Third Department in People v Caden N.

The key issues were:

• Whether, at the time of trial, People v Caden N. clearly entitled a VTL § 1192 (4)/(4-a) defendant to a “heightened” instruction equating drug “impairment” with alcohol “intoxication” under People v Cruz.
• Whether counsel’s failure to seek that charge was an “egregious and prejudicial” error under New York’s “meaningful representation” standard.

2. Summary of the Opinion

The Court of Appeals affirmed. It held that trial counsel was not ineffective for failing to request a Caden N.-based instruction in this VTL § 1192 prosecution because, at the time of trial: (1) Caden N. was “carefully limited” to the vehicular manslaughter statutory scheme, (2) no New York appellate court had extended Caden N. to VTL § 1192 (4)/(4-a) prosecutions, and (3) the model Criminal Jury Instructions in effect still tracked People v Cruz, and the Court has long advised reliance on them.

Judge Rivera concurred in the result but offered different reasoning: counsel should have requested the more favorable Caden N. instruction because it was binding Third Department precedent and its penalty-scheme logic “squarely applied.” However, counsel still was not constitutionally ineffective because the Appellate Division on direct appeal held defendant was not entitled to that instruction—defeating the necessary showing of prejudice/fair-trial deprivation.

3. Analysis

3.1 Precedents Cited (and How They Shaped the Decision)

People v Cruz (48 NY2d 419 [1979], appeal dismissed 446 US 901 [1980])

Cruz supplies the foundational vocabulary: “impaired” (for alcohol) means the defendant has impaired “to any extent” the abilities expected of a reasonable and prudent driver, while “intoxicated” denotes incapacity to employ those abilities. Ambrosio treats Cruz as the baseline from which both the model instructions and the later Caden N. debate depart.

People v Caden N. (189 AD3d 84 [3d Dept 2020], lv denied 36 NY3d 1050 [2021])

Caden N. held—“in the context of assessing whether a person has committed the crime of vehicular manslaughter in the second degree”— that drug “impairment” should match the alcohol “intoxication” standard from Cruz (incapable of employing expected abilities).

In Ambrosio, the majority emphasizes the limited scope of Caden N. and the absence, at the time of trial, of any appellate extension of that reasoning to standalone VTL § 1192 (4)/(4-a) charges. That framing is crucial to the ineffective-assistance outcome: an attorney is not constitutionally deficient for failing to press a favorable instruction when entitlement to it is unsettled and would require a novel extension.

People v Watkins (42 NY3d 635 [2024], cert denied 145 S Ct 459 [2024]) and People v Clark (28 NY3d 556 [2016])

These cases anchor the modern articulation of New York’s ineffective assistance standard: the defendant must show the lack of “meaningful representation” resulting in a deprivation of a fair trial. Watkins is repeatedly used in Ambrosio to frame (1) the objective “whole representation” assessment and (2) the “rare” nature of single-error reversals.

People v Benevento (91 NY2d 708 [1998]) and People v Baldi (54 NY2d 137 [1981])

Benevento and Baldi support the Court’s insistence that counsel is judged in context—“in light of the evidence, the law, and the circumstances at the time of the representation”—and with emphasis on process fairness rather than outcome determinism.

People v Caban (5 NY3d 143 [2005]) and People v Turner (5 NY3d 476 [2005])

These decisions supply the “single error” framework: one mistake can constitute ineffective assistance only if it is sufficiently “egregious and prejudicial,” often where counsel fails to raise a “clear-cut and completely dispositive” defense. Ambrosio uses this line to conclude that the omitted Caden N. request was not clear-cut or dispositive at the time.

People v Saenger (39 NY3d 433 [2023]) and People v Borrell (12 NY3d 365 [2009], rearg denied 12 NY3d 365 [2009])

These cases underscore the narrowness of the “rare” category and reinforce that counsel is not ineffective for failing to pursue an issue of “uncertain efficacy.” The majority treats the attempted extension of Caden N. as precisely such an uncertain proposition.

People v Hayward (42 NY3d 753 [2024]) and People v Keschner (25 NY3d 704 [2015])

Hayward supplies the key principle applied to this record: counsel is not ineffective when success would have depended on the resolution of novel questions or when there was no clear appellate authority supporting the argument at the time of trial. Keschner reinforces that the Court may reject ineffective-assistance claims where the omitted contention would not have been a “winning argument” leading to reversal.

People v J.L. (36 NY3d 112 [2020]) and People v Cubino (88 NY2d 998 [1996])

These cases are cited for a practical, institutional point: the Court has advised counsel and trial courts to rely on the Criminal Jury Instructions. That matters in Ambrosio because the then-current CJI for VTL § 1192 (4)/(4-a) paralleled Cruz, and the CJI Committee did not revise it to address Caden N. until after trial.

People v Thompson (21 NY3d 555 [2013]) and People v Blake (24 NY3d 78 [2014])

These decisions provide the Court’s phrasing that even if counsel’s choice was a “mistake,” it must be the kind of “egregious and prejudicial” error that deprives the defendant of the constitutional right to counsel. Ambrosio applies that restraint to the charge-request omission.

Cases prominent in the concurrence: Strickland v Washington (466 US 668 [1984]), People v Droz (39 NY2d 457 [1976]), and People v Litto (8 NY3d 692 [2007])

Judge Rivera’s concurrence draws a sharper duty-to-research theme from People v Droz and frames prejudice through both the federal baseline (Strickland v Washington) and New York’s “meaningful representation” standard. The concurrence also leans on the penalty-scheme discussion in People v Litto (quoted by Caden N.) to argue Caden N. should have been invoked at trial as a more favorable, binding Departmental rule.

People v Rossi (163 AD2d 660 [3d Dept 1990], lv denied 76 NY2d 943 [1990])

Rossi is relevant historically because Caden N. overruled it to the extent it could be read to apply Cruz’s “impaired to any extent” standard to vehicular manslaughter involving drugs. In Ambrosio, that lineage illustrates that the law in this area had been in flux—supporting the majority’s “unsettled/novel extension” characterization.

3.2 Legal Reasoning

The majority’s reasoning proceeds in three moves:

1. Identify the claimed omission: counsel did not seek a jury instruction defining drug impairment under the more demanding “incapable” standard used by Caden N. (importing Cruz’s “intoxication” language), and instead requested the “standard” instruction then in the CJI—tracking Cruz’s “impaired to any extent” formulation.
2. Ask whether the omitted argument was clearly available at the time: the Court stresses that Caden N. explicitly confined its holding to vehicular manslaughter and, when defendant was tried, no appellate authority had extended Caden N. to VTL § 1192 (4)/(4-a) prosecutions. This invokes the doctrinal guardrails in People v Hayward and the “rare case” rubric of People v Turner/People v Watkins.
3. Measure egregiousness and prejudice/fairness: given the uncertain entitlement, counsel’s reliance on model CJIs—especially where the Court has endorsed that reliance (People v J.L.; People v Cubino)—was not the sort of objectively unreasonable, outcome-skewing failure that deprives a defendant of a fair trial.

Judge Rivera’s concurrence accepts a different premise—Caden N. was “binding precedent” in the Department and its penalty-scheme logic applied—yet reaches the same result by focusing on the posture of the case: the Appellate Division on direct appeal concluded defendant was not entitled to the heightened instruction, so defendant could not establish the requisite prejudice (federal) or unfairness of the process (state).

3.3 Impact

Ambrosio is most significant in the ineffective-assistance lane rather than as a standalone exposition of VTL § 1192’s substantive impairment standard. Its practical effects include:

• Reinforcing “no-IAC-for-unsettled-extension” doctrine: the case underscores that defense counsel’s failure to press a more favorable theory will not, without more, constitute ineffective assistance when success would have required extending precedent beyond its stated scope.
• Confirming protective value of following CJIs: where counsel requests the model charge and the model charge reflects existing Court of Appeals authority (here, People v Cruz), Ambrosio makes it harder to repackage that decision as constitutional ineffectiveness.
• Limiting retroactive leverage from short-lived appellate innovations: even where an intermediate appellate decision (like Caden N.) suggests a defendant-favorable doctrinal shift, Ambrosio signals that counsel’s failure to anticipate broader adoption will rarely be treated as an “egregious” constitutional error.
• Interaction with subsequent substantive clarification: the majority’s reference to People v Dondorfer indicates that, going forward, litigants should look to Dondorfer for the governing substantive rule about uniform impairment standards across VTL § 1192, while Ambrosio addresses what happens when counsel did not anticipate (or attempt to extend) an unsettled approach before that clarification.

4. Complex Concepts Simplified

“Impaired” vs. “Intoxicated” (from People v Cruz)

• Impaired: abilities are diminished to any extent—even modestly—below what a reasonable, prudent driver should have.
• Intoxicated: the person is incapable of employing the abilities a reasonable, prudent driver should have.

What People v Caden N. did (and why it mattered here)

Caden N. (in vehicular manslaughter litigation) treated “impairment by drugs” as requiring the same degree of functional compromise as alcohol “intoxication,” largely to avoid penalty inconsistency where deaths caused by similarly affected drivers would be punished differently depending on substance. Defendant in Ambrosio argued that logic should carry over to ordinary VTL § 1192 (4)/(4-a) prosecutions.

New York’s ineffective-assistance standard (“meaningful representation”)

Under People v Watkins, People v Benevento, and People v Baldi, the inquiry focuses on whether counsel’s overall performance deprived defendant of a fair trial, viewed objectively at the time—not whether a different tactic might have improved the odds. A single omission warrants relief only in rare situations where the missed issue was clear, dispositive, and its absence undermined trial fairness.

5. Conclusion

People v Ambrosio holds that trial counsel was not constitutionally ineffective for failing to request a jury instruction extending People v Caden N.’s heightened drug-impairment standard to VTL § 1192 (4) and (4-a) prosecutions, where that extension was unsettled at the time, no appellate authority had adopted it, and counsel relied on then-current model instructions tracking People v Cruz. The decision strengthens the principle that ineffective-assistance claims cannot be built from hindsight-driven expectations that counsel should have pursued novel or uncertain doctrinal expansions—particularly when prevailing CJIs reflected existing Court of Appeals authority.