Police Warn High-Risk PREDATOR Is Back on Streets

crime scene tape with police officers in a park — Photo: Steve Sanchez Photos / Shutterstock

When authorities warn that someone “poses a significant risk to the community, especially children,” yet that person is lawfully released under tight conditions, you are looking at the core tension in modern public safety: the line where completed punishment ends and preventive control begins.

The Short Version

Durham Regional Police publicly designated Keith Theodore Constantin a high-risk offender and warned he posed a “significant risk,” especially to children.
He was released subject to unusually stringent community conditions: GPS monitoring, curfew, weekly police reporting, broad child-contact and location bans, and prohibitions on weapons, alcohol, drugs, and internet use.
Local mayors and police framed the release as a supervised, high-control measure; a Conservative MP used the case to attack Liberal “soft-on-crime” policy.
Canada’s legal framework separates punishment for past crimes from preventive recognizances designed to manage future risk; breaches of such orders are themselves criminal offences.

What was publicly established: the warning, the release, the controls

Durham Regional Police issued a public safety alert identifying Keith Theodore Constantin, age 48, as a high-risk offender whose presence in the community posed a significant risk, “especially children,” and catalogued a criminal history including sexual assault, sexual assault with a weapon, assault with a weapon, assault, robbery, possession of explosives, uttering threats, and multiple probation violations. Local reporting reproduced the police characterization and the decision to release him under supervision in Durham Region. The police described a robust suite of conditions: electronic and GPS monitoring, a 7 p.m.–7 a.m. curfew, weekly in-person reporting, prohibitions on contact with anyone under 16 and on attending places where children gather (schools, playgrounds, daycares, swimming areas, malls), and bans on weapons, alcohol, drugs, and internet use or internet-capable devices.

Municipal leaders amplified the alert. Oshawa’s mayor acknowledged the risk designation and emphasized collaboration with police. Whitby’s mayor clarified that Constantin had completed a court-imposed sentence—not a bail or temporary absence—and expressed confidence that DRPS could enforce the court-ordered conditions.

How Canadian law manages risk after sentence expiry

Canadian criminal law draws a hard line between punishment for completed crimes and preventive measures aimed at future risk. When a sentence ends, continued confinement is not the default; the state must rely on specific legal tools designed for risk management. One such mechanism is a preventive recognizance under Criminal Code section 810.2—an order a court can impose when there are reasonable grounds to fear that a person will commit a serious personal injury offence. The recognizance can carry conditions limiting movement, contacts, places, and conduct, tailored to the perceived risk profile.

These recognizances are not mere suggestions: breaching them is a standalone criminal offence under section 811, enabling police to arrest and Crown counsel to prosecute non-compliance as an immediate public safety response, rather than waiting for a new substantive offence to occur. In practice, police offender management units use curfews, GPS, frequent check-ins, and place and association restrictions to create enforceable guardrails around an individual who is living in the community but flagged as high risk.

Why the release and the rhetoric collide

The Constantin case triggered two competing frames that often talk past one another. Police and local officials presented a control regime: completed sentence, intensive supervision, rapid breach consequences, and a clear public warning. That frame treats community release with stringent conditions as a public safety tool rather than permissiveness. Critics, by contrast, highlight the danger signals—serious past violence, explicit risks to children, prior supervision breaches—and argue that any release under such a profile reflects systemic leniency that places the public, especially minors, at unacceptable risk. A sitting Conservative MP in the region used precisely that language, tying the case to “catch-and-release” and “soft-on-crime” policy claims against Liberal leadership.

Both frames have anchors in fact. The conditions here were unusually burdensome and specific, consistent with high-risk management best practices—GPS tethering, curfews, and categorical bans on child contact and child-centered spaces are the backbone of risk containment in the community. Yet the risk signals are not hypothetical; DRPS publicly enumerated a serious, multi-offence history and explicitly warned of a significant risk to children. The public-policy argument turns on which cost you are prepared to accept: the civil liberty limits of long-tail controls, or the residual risk that remains even after strong conditions are applied.

How these orders are supposed to work in real life

Mechanically, preventive recognizances rely on three levers. First, surveillance: GPS geofences and curfews generate tamper or zone alerts; weekly reporting and unannounced checks keep direct contact pressure on the subject. Second, exclusion zones and association bans reduce opportunity—particularly vital in child-protection cases where place-based risk is high. Third, swift sanction: the breach offence under section 811 permits prompt arrest and charge for any condition violation, re-establishing immediate custody if the risk window widens.

The effectiveness of electronic monitoring and strict conditions is mixed in the academic literature; some studies find limited or context-dependent impacts on long-term recidivism, which is not the same question as short-term incapacitation or rapid breach detection. In practice, the operational aim is not to “cure” risk but to compress the time between risky behavior and intervention. Properly configured conditions, coupled with active supervision by a dedicated offender management unit—as DRPS said it would deploy—are the practical expression of that strategy.

Politics aside: what this case can and cannot tell us

As a matter of record, DRPS publicly warned that Constantin is high risk and subject to a dense net of conditions; local mayors endorsed the enforcement posture; a Conservative MP cast the release as evidence of Liberal mismanagement. What the public documentation does not do is attribute the release to a particular ministerial directive or unique statutory change traceable to one party. The lawful route for managing risk after sentence completion is the Criminal Code’s preventive recognizance scheme and related supervision practices—longstanding tools that predate today’s rhetoric and bind courts, police, Crowns, and communities to the same balancing act between liberty and security.

One further boundary deserves precision. A public safety alert does not itself establish new offending. The available reporting documents the release, the rationale for the restrictions, and the conditions themselves; it does not identify post-release crimes or breaches. That omission neither validates nor invalidates the risk designation; it simply marks the limit of what is publicly reported to date.

The enduring question for communities

Cases like this are not anomalies; they are the predictable product of a legal system that ends punishment at sentence expiry and manages ongoing danger with civilian life under constraint. When police describe a subject as a significant risk to children and then surround that person with GPS, curfews, exclusion zones, and immediate-breach consequences, they are using the tools Parliament has provided to hold two obligations at once: protect the public and respect the boundary that separates completed punishment from indefinite detention. Whether that trade is tight enough—or too tight—is the debate. The facts here are clear about the risk, the release, and the regime built to contain it; the judgment about sufficiency rests with voters who must decide how they want their laws to draw that line.