Is Ignorance Bliss, or Is It Required?

Learning from failure, the fear of the public record, and what law enforcement training can borrow from naval aviation

TL;DR

• When I show chiefs what AI-driven training analysis can do, the most common response is not about accuracy. It is, "Is this information FOIA-able?"
• The best learning happens when the learner is failing some of the time. Naval aviation built a system around that. The surface Navy did not.
• "Will someone get this record?" is not one question. It is three: public disclosure, civil discovery, and use at trial.
• Ignorance is not a safe harbor. The way through is to decide in advance what you will do with the answer.

"I don't think I want to know this much information."

I hear this phrase routinely from law enforcement executives across the country, from the chief of a 10-person department to the chief of an agency with more than 3,000 sworn officers.

It usually comes up when I show them what AI-driven training analysis can do. A tool can now review hundreds, if not thousands, of training encounters and show a department where each officer needs help: the objectives they meet, the ones they miss, and the mistakes they keep making. The most common response is not about accuracy. It is, "Is this information FOIA-able?" or "I do not want this much analysis done on my department's officers."

I understand the mistrust of AI and its tendency to make errors. Auditing, recordings, transcripts, and a human checking the analysis can address that. But these conversations go beyond whether the AI is accurate. The concern is what happens if it is accurate, and now there is a record of something the department needs to address.

A chief should want to find that gap before an officer has to make the decision on the street. Yet if finding it and documenting it feels like creating a problem, there is a reason to avoid looking too closely. That is the problem this post is about.

One note on scope. This is not legal advice, and public-records and discovery rules vary by state. Every legal point here should be run past agency counsel.

The Navy Gave Me a Place to Fail

I started comparing these conversations to my time as a Navy aviation engineer. I caused the loss of hundreds of millions of dollars' worth of naval assets, and thousands of simulated personnel, all in a simulated environment, because that is where I trained. That environment let me set up scenarios destined to fail, then analyze my performance and my ability to adapt. My supervisors looked at every error and debriefed it with me.

An old Navy chief used to tell me, "It's all about the 6 Ps, AJ. Always remember your 6 Ps."

"What are the 6 Ps, Chief?" I asked.

"Proper Preparedness Prevents Piss Poor Performance."

And yes, I failed in real events too. Luckily no lives were lost, but we did lose valuable test assets because of my decisions. That incident gave me a reason to improve even more, and it became a training scenario for others.

Athletes already know this. Receivers practice basic route running. Kickers practice chip-shot field goals. When they fail, they practice even more. There is research behind it too. In 2019, a team led by Robert Wilson found that the learning rate is highest when the error rate sits at about 15.87 percent, which they called the "85 percent rule." A training program in which nobody ever fails is either too easy or not honest.

So why is it that in a profession where officers can strip citizens of their liberty by making an arrest, we have officers and executives afraid of failing in practice?

Two Communities, One Navy

The clearest example I know of what culture does to learning sits inside a single service branch.

I was around for the 2017 ship collisions. USS Fitzgerald collided with a merchant ship on June 17, 2017, killing seven sailors. USS John S. McCain collided with a tanker on August 21, 2017, killing ten. The Navy's own review called both accidents preventable.

Both communities are the same Navy. Their cultures toward rest and training are very different.

Aviation. Naval aviation runs on hard limits. The squadron-level instructions I found for Navy flight training, for example, require 12 hours of continuous crew rest before a crew day begins. Training is not what you do after the real work. It is part of the real work.

Surface. The surface force has a very different record. The Navy's 2017 comprehensive review described forward-deployed ships operating with lapsed training certifications, some for as long as 22 months in the case of Fitzgerald. Navy Times reported that, unlike aviation and submarines, surface ships lacked mandated sleep guidance. As recently as 2023, the Government Accountability Office reported a surface-force average of 5.25 hours of sleep, against a standard of at least 7.5 hours available to sleep each day.

A former surface officer put it to USNI News this way: the pride of the surface fleet is the ability to say yes.

The contrast with aviation is not an accident. Aviation learned from its losses. USNI News reported that almost 12,000 aircraft were lost between 1949 and 1988, and that this drove the standardization and discipline that define naval aviation today. Aviation turned failure into procedure.

Same Navy. Same oath. One community built a system to learn from failure, and the other built a culture that was proud of not needing one.

Logic Is the Underdog

You can make the logical case all day. Failure in practice is cheap, failure on the street is not, and records of improvement protect everyone. The case is sound. It also loses to the limbic brain, the part of us that reads "record of a mistake" as "threat."

People fear failure. Departments avoid it by not looking, officers avoid it by not attempting the hard scenario, and instructors avoid it by grading generously. Every one of those choices is rational from inside the fear. Every one of them leaves the gap in place.

One police executive told me: "There is a lot more downside to training records than there is upside. Training must be a safe space where we can fail away from cameras, recordings, and AI, the more we can learn." I take that seriously. He is not arguing against learning. He is describing what the fear costs him.

To get past the emotional brain, every person in the chain needs an answer to one question: What's In It For Me? Telling people to be brave is not an answer. Each audience needs its own.

My own experience is the proof of concept. My records were pulled after a real failure. Because they showed that I had practiced, tried to improve, and followed procedure, they helped me instead of hurting me. The record was not the danger. The absence of a record of improvement would have been.

"Is This FOIA-able?" Is Really Three Questions

When a chief asks whether something is FOIA-able, the law that usually applies to a local department is the state's public-records law. Federal FOIA covers federal agencies. But "will someone get this record?" is not one question. It is three, and they have different answers:

1. Public disclosure. Can a member of the public or a reporter obtain it through a records request? That depends on the state statute and its exemptions.
2. Civil discovery. Can a plaintiff's lawyer obtain it in a lawsuit? In federal civil cases, relevant and proportional information can be discoverable even if it would not be admissible at trial.
3. Use at trial. Can it be put in front of a jury? That is an evidence question, separate from both of the above.

A public-records exemption alone does not answer the other two. A department cannot create privacy by calling a record internal. Courts want a defined category, not a label. Protection has to come from a legislature or an agency rule that deliberately builds it.

Aviation has done exactly that, in several ways.

• Military safety investigations. Courts have long protected confidential statements given to military air-crash safety investigators, because witnesses speak more freely when they know the statements are used for accident prevention.
• NASA's Aviation Safety Reporting System. ASRS removes personal and organizational names. NASA reports more than a million submissions with no documented breach of reporter identity.
• Flight data and voluntary programs. Under 14 CFR Part 193, the FAA does not disclose protected voluntarily provided safety information in response to FOIA requests.

Every one of these has the same features. The protected category is defined in advance. Identity is stripped or access is controlled. There are explicit carve-outs for accidents, crimes and ongoing hazards. And a legislature or agency created it on purpose. The honest position is that aviation proves the concept works. It does not prove that law enforcement can copy it without changes.

The courts already give departments a reason to take training seriously. In City of Canton v. Harris (1989), the Supreme Court held that inadequate training can support municipal liability when the failure amounts to deliberate indifference to constitutional rights and causes the injury.

So ignorance is not a safe harbor. A department that sees a serious gap and does nothing has a different problem from an officer who made a mistake while learning.

Decide Before You Look

Set FOIA aside. Suppose the protections exist and the records are safe. Analysis at this scale will show things no department has seen before, and some of it may be things a department would prefer to leave hidden. Is that what we want?

We have opened this box before. Early warning systems are not new. What changes with AI is speed. Analysis of training encounters can show a skill gap long before it becomes a complaint or a use-of-force review, in something close to real time. These are the risk areas I would want leaders to think about before they turn the analysis on, not after.

• The training program may be the finding. Analysis can show that scenarios are unrealistic, instructors grade inconsistently, or standards are unclear.
• The schedule may be the finding. If performance drops with long shifts, heavy overtime or court time, the data may say the cause is a staffing or scheduling decision, not the officer. That is the surface Navy's sleep problem in a different uniform.
• Who holds the data. A protection means little if the records sit in a vendor's system the department cannot control or export. That is the subject of my earlier piece on vendor lock-in.

The way through is not to avoid looking. The gap exists whether or not it is recorded. The officer still has to make the decision, and the citizen still has to live with the result. The way through is to decide in advance what you will do with the answer.

So the question stands. Is ignorance bliss? Only until the call comes in. After that, it is the most expensive thing a department can own.

Who Should Lead This

A single department cannot solve this alone. No chief wants to be the first to put practice records on the line, and no software company can promise a legal protection by changing a setting. This needs the associations.

ILEETA, IADLEST and IACP should lead the conversation. Put instructors, officers, agency attorneys, civil-rights representatives and legislators in the same room. Work out what should be protected, who can still see it, and what a department must do when it finds a gap.

Departments do not have to wait. Tell officers when they are practicing and when they are being formally evaluated. Explain who sees the results and what happens if an instructor identifies a serious concern. Give officers a way to challenge an incorrect score. Assign someone to review each confirmed gap, provide coaching, and check that the officer can demonstrate the objective afterward.

The Navy gave me a place to make mistakes and people who would sit down and go through every one with me. We recorded what happened, learned from it, and practiced again. Officers deserve the same.

I keep coming back to those conversations. "I don't think I want to know this much information." If that is where we are, we have work to do, because the officer still has to go out and make the decision, and the weakness we were afraid to look at is still there.

Proper Preparedness Prevents Piss Poor Performance. Give officers a place to prepare.