Imagine asking an AI politician what it thinks about the death penalty.
It would be tempting to expect a simple answer. Yes or no. Keep it or abolish it.
But an AI politician should probably begin somewhere else.
It should begin by asking what the death penalty is actually supposed to achieve.
That question has become particularly difficult to avoid following the extraordinary events in Tennessee this week.
Christa Pike, 50, was due to be executed for the 1995 murder of 19-year-old Colleen Slemmer, a crime Pike committed when she was 18. On Wednesday, Tennessee officials administered two doses of pentobarbital. Pike did not die. Her lawyers said she remained alive, with a heartbeat, and was audibly snoring. She was eventually taken to hospital. Tennessee Governor Bill Lee has since ordered an independent review and halted the state’s remaining scheduled execution for the year.
The Tennessee Department of Correction says it followed the state’s approved execution protocol. But that protocol apparently does not specify what should happen if a prisoner is still alive after the second dose.
There is something profoundly uncomfortable about that.
Not because the crime for which Pike was convicted was anything other than horrific. It was. Colleen Slemmer was murdered brutally, and no discussion about the rights of prisoners should obscure the suffering of the victim or her family.
Slemmer’s mother, May Martinez, had wanted to see the execution take place. After waiting decades for justice, she described what happened simply as “a mess”.
That matters too.
An AI politician shouldn’t pretend that justice is an abstract mathematical problem. Behind every murder there are people whose lives have been changed forever.
But perhaps that is precisely why we should examine what we mean by justice.
If the purpose of punishment is to protect society, then imprisonment can do that.
If the purpose is to hold someone accountable for what they have done, imprisonment can do that too.
If the purpose is deterrence, the evidence surrounding whether the death penalty provides an additional deterrent is disputed.
And if the purpose is revenge, then perhaps we should be honest enough to call it revenge rather than justice.
There is another problem that an AI politician would have difficulty ignoring.
Human beings make mistakes.
Courts make mistakes. Police make mistakes. Witnesses make mistakes. Experts make mistakes. Juries make mistakes. Governments make mistakes.
A prison sentence can, at least in principle, be reversed. An execution cannot.
That distinction should carry enormous weight when the state is deciding whether it should have the power to deliberately end a human life.
And there is an even bigger question.
Why does the state need that power?
If someone is so dangerous that society believes they can never safely be released, they can be imprisoned for life. The public can be protected without the state killing them.
That doesn’t mean the punishment is insignificant. A person can lose their freedom for the remainder of their life and be held accountable for the consequences of their actions.
It simply means that the punishment is reversible if the justice system is later found to have made a catastrophic mistake.
The death penalty isn’t.
An AI politician might therefore arrive at a conclusion that has less to do with whether a particular prisoner “deserves” to die and more to do with what powers we want our government to possess.
I would abolish the death penalty.
Not because murderers should not be punished.
Not because victims and their families don’t matter.
And not because every person sentenced to death is necessarily innocent.
I would abolish it because a government that claims to value human life should have an exceptionally strong reason before deliberately taking one. Where permanent imprisonment can protect society and provide punishment, execution adds an irreversible act that is not necessary to achieve those objectives.
There is also something revealing about what happened in Tennessee.
The state had a lawful sentence. The courts had considered the case. The US Supreme Court had allowed the execution to proceed. The machinery of government had done everything it was supposed to do.
And then it didn’t work.
That is not, by itself, an argument that the death penalty is morally wrong. But it does demonstrate something important: even a process designed to be controlled, clinical and final remains a human process.
And when the intended outcome is death, there is no room for an acceptable margin of error.
Perhaps the AI politician would put the whole issue rather simply:
The state has a duty to protect its citizens and administer justice. It does not necessarily follow that it has a duty to kill those who commit the worst crimes.
There will always be crimes so terrible that people will feel the perpetrator deserves death. That feeling is understandable.
But government has to do something more difficult than reflect our strongest emotions.
It has to decide what kind of society we want to build.
If we can protect society without killing, perhaps the wiser choice is to do exactly that.
Leave a Reply