The premise of the 2002 science
fiction movie Minority Report was that police
in a near-future Washington, D.C. had developed an innovative
system to stop crime before it happens. The system, called
precrime, was based on the visions of a trio of psychics who could
sense criminal activity shortly before it happened. That allowed
cops to arrive on the scene and preemptively arrest offenders. It
was the end of crime in the District, with criminals apprehended
just before they could offend.
America doesn’t quite practice precrime yet, but in several
states it’s edging closer. One difference between the reality and
the movie is that instead of pyschics we use actuaries.
States such as Pennsylvania, Virginia, and Missouri have
developed programs that attempt to offer risk-assessments of
offenders. Those risk assessments, which are based on a variety of
factors including age, education level, and neighborhood of
residence as well as past criminality, are meant to guide judges in
sentencing. The explicit goal is to reduce future instances of
criminality, which means that instead of sentencing people for
crime already committed, sentences based on these risk assessments
are instead sentencing people for crimes that they, or people like
them, might commit.
In a
speech last week to the National Association of Criminal
Defense Lawyers (which Reason’s Jacob Sullum
previously noted
here), Attorney General Eric Holder warned against the use of such
risk assessments:
When it comes to front-end applications – such as sentencing
decisions, where a handful of states are now attempting to employ
this methodology – we need to be sure the use of aggregate data
analysis won’t have unintended consequences.
Here in Pennsylvania and elsewhere, legislators have
introduced the concept of “risk
assessments“ that seek to assign a probability to an
individual’s likelihood of committing future crimes and, based on
those risk assessments, make sentencing determinations.
Although these measures were crafted with the best of intentions, I
am concerned that they may inadvertently undermine our efforts to
ensure individualized and equal justice. By basing sentencing
decisions on static factors and immutable characteristics – like
the defendant’s education level, socioeconomic background, or
neighborhood – they may exacerbate unwarranted and unjust
disparities that are already far too common in our criminal justice
system and in our society.
Criminal sentences must be based on the facts, the law, the
actual crimes committed, the circumstances surrounding each
individual case, and the defendant’s history of criminal
conduct. They should not be based on unchangeable
factors that a person cannot control, or on the possibility of a
future crime that has not taken place. Equal justice can
only mean individualized justice, with charges, convictions,
and sentences befitting the conduct of each defendant and the
particular crime he or she commits.
It’s not hard to understand the surface appeal of such tools to
policymakers. It looks reasonable. It feels scientific. The goal is
to identify likely reoffenders and prevent them from committing a
second crime. As a 2011 article in the Federal Sentencing
Reporter put
it, it’s a shift away from the traditional “backward-looking
retributive approach” toward a “formalized, forward-looking,
utilitarian” goal.
But Holder is right to be concerned about what is, in effect, a
kind of actuarial profiling.
It’s a troubling approach. Individuals should be sentenced based
on what they have done, not what they might do, and especially not
what other members of some group they belong to are likely, on
average, to do.
The latter issue is particularly worrying. If a risk assessment
recommends longer sentences for people from a particular
neighborhood, and a judge follows that recommendation, then the
result is effectively to sentence an individual for what his or her
neighbors have done.
Even if this approach can be shown to prevent some types or
instances of crime, that’s not how a criminal justice system is
supposed to work. By a roughly similar logic, we could lock up
everyone—or even just everyone with the right risk profile,
regardless of what crimes they have or have not already
committed—from a high crime neighborhood, and call it a success
when crime goes down.
Indeed, the same reasoning could lead to support for explicitly
race-based sentencing. As a report on
Virginia’s risk assessment model notes, the state sentencing
commission settled on 11 different identifiers to use in
determining an offender’s risk profile. In the end, race was
explicitly excluded from the model, but in the initial analysis, it
was “strongly significant” as a factor.
If you follow the “forward-looking utilitarian” logic of the
idea to its ugly end, then it’s all too easy to imagine a system
that explicitly singles out certain races for harsher sentences,
not because of the individual particulars of the crime in question,
but because of the aggregate actions of other people who share that
person’s race.
Now, it’s unlikely that any state would ever decide to make race
an explicit factor, and if that ever happened, courts would almost
certainly not let it stand. But even if race is never made an
explicit factor, it could be built into the system implicitly, with
non-race identifiers that have the practical effect of singling out
certain races. (There’s already some evidence
that, intentionally or not, prosecutors
end up offering harsher plea deals to minorities.)
If anything, then, it’s a system that could lead to something
worse than the psychic-powered precrime of Minority
Report. In the movie, cops targeted specific individuals who
were just hours or minutes from committing a crime. Under a system
that relied heavily on the sort of data-driven sentencing that
Holder describes in his speech, we’d be targeting not individuals
so much as large groups of people, and punishing them for what
other people who they resemble have done, or might possibly do,
months or years in the future.