Trust Must Be Earned

And AI is NOT earning it.

It all started with a simple question intended to help resolve a scheduling conflict. "Is the 'Autumn in the Mountains Car Show' always on the last weekend of September?" This info is not directly on the organizer's web site, but past event dates could be manually searched to see if history supports the assertion. But manually? Isn't that what AI-powered smart Giggles are for? And so it was Giggled, and Gemini responded.

And Gemini is wrong, at least in the particulars. 

History indeed indicates the event is not always held on the last weekend in September and isn't consistently even the fourth Saturday. Fair enough. Top level answer is correct if you can trust AI. But can you? Well, AI suggests the answer is no, in the very answer given above. The event was held on 9/26/26, but this past Saturday is not the "second-to-last" Saturday (whatever that really means), and is, in fact, the fourth and last Saturday of September 2026. It must take a silicon based mind to think that is "second-to-last."

So...do you trust AI? Do you trust AI to be the ubiquitous, all seeing, all knowing overlord of your, and your family's, life? If so, you are either a city bureaucrat, or an elected official who enjoys private (not secret...no, no, not secret) get-togethers with corporate bigwigs. Otherwise, asking simple questions, either of these folks, or AI itself, will destroy any trust you may have had. Wouldn't it? Hasn't it?

So why have we allowed these folks to replace organic intelligence with the artificial variety, especially in the domain of public safety? If, in your opinion, this is the wrong decision, you may want to consider replacing these folks who clearly disagree.

Gimme Shelter

If AI companies are anything they are smart. And they are learning. They are learning a lot from what has recently happened to social media operations, including some of their very own. Once high-flying, they've had their wings clipped, and some jurisdictions have an outright ban on social media for minors. There have been quite a few, very large, court awards that are costing enormous amounts of money and impacting their operations at additional cost. It is also a PR nightmare that may bleed over to AI operations.

They do not want to endure rinse and repeat with their AI operations. 

That's why they want to put on the brakes. Not because they really want to, but because they want to avoid a future where they will be crushed by fines and overbearing restrictions. So what they want to do is slow their roll until they can get regulations in place. Regulations they can live with, and while somewhat restrictive, the real goal is to offload any, perhaps all, future liability. They want regulations, crafted by their lobbyists, that they can easily follow, and if they do, any Bad Things are because of the politicians and regulators. After all, they were in compliance with all laws, rules and regulations. 

They also know that once these regulations are passed, politicians will have their name on them, politicians who are now politically motivated to defend the AI industry rather than attack it. How do the AI firms know this will work? Because that is exactly what Flock has done.

These actions will not prevent all future lawsuits but it will provide shelter from those enormous jury awards. 

Is It Misinformation?

Is it misinformation if it isn't intentional? If it is intentional is that disinformation? Can incompetence, say in reporting, be mere misinformation? 

Here's what we have.

There are a few things wrong with this. First, Dunwoody doesn't have a fire department, so those cannot be Dunwoody firefighters. They are (maybe?) DeKalb County firefighters, and thank god for that. Then, as it turns out, or so history tells us, 2977 people did not die in the twin towers on 1 Sept 2001. However 2977 did die in the terrorists' attack on the twin towers ten days after that date.

You have to wonder how much time must pass for events to become history, and historical facts that set the record straight when someone prevaricates. A week? A month? Years? Decades? Perhaps time is not the measure. Perhaps it is acceptance. Entry into the zeitgeist. Sometimes folks forget. Some never knew. Approximately 103 million people in America today were born after 9/11. It was never part of their existence. Over half of Americans were born after that last helicopter lifted off in Saigon. Is 9/11 headed to the same place as Vietnam? Some sepia stained story told by increasingly irrelevant old farts? Is it already there?

Nothing More Conservative...

...than a liberal.

Resist!

The very essence of "resist" is fight the change. Sounds like the conservative affection for preservation. Yet now this is coming from the self-professed radical left. Or, nearby, those that refuse to disclose openly, pretending it isn't the most open secret in daVille. Everyone sees the true colors. 

Ironically these are the folks trying their dead level best to create an overarching surveillance state. That seems like something you'd expect from the law-and-order camp. You know, the folks the lefties call nut-jobs. But it is the "abolish cash bail" folks who want everyone continuously surveilled by AI. 

And then, these same proponents of AI surveillance are sending their army of cultivated protestors out to condemn any and all data centers. The same data centers used by AI systems. What is their long game? Let China completely dominate the data center market? So, their plan is to use AI for heavy-handed surveillance of all of us, but to also have all that data hosted in China. Or would they hide that behind "unintended consequences" which is the most common outcome for any government activity? Ooops.

That's not the real problem. The real problem is these fools were elected, and they probably will be again.

Let AI Do It

It is time for a start.Well past time, in fact. Since before this city was formed there was a sincere dedication to opacity. Did you ever see the officially published Task Force reports? No? Well, there's a reason for that. They, many of whom fancied themselves the powers-to-be, had no intention of anything resembling full disclosure. So no surprise that this city keeps all the juicy bits of information behind an enormously expensive pay wall. Might as well be locked up in Fort Knox. 

We can do better.

And it is easy to do because we can let AI do it for us. And we can start with the easy button, addressing opacity with a "service" that is a self-proclaimed leader in the use of AI. That's right. Let's use Flock on Flock. Start simple. Post a daily report of all the license plate numbers, with state, captured inside the city limits. Stretch goal: list the number of times that day the plate was surveilled. Then any of us, we the little people, would have automatic access to public data. It would be fun just to sift thru the grift looking for out of state plates, your own plate, your spouse's plate, your neighbors...this could replace death scrolling. 

So how will the city say no?

They may pay-wall it or just claim it is too expensive. Really? This truly would be trivial for damn near any half-assed IT geek, and frankly requires no AI. There are other fun things that AI could be used for, but this is no heavy lift. But, it is critically important not to strain the brains of the city bureaucrats. Keep it stupid. Simple? There is one thing they will probably bring up as an excuse for not doing this: privacy. Kinda sucks the air out of the room, doesn't it? They've justified their overarching surveillance because this is all in public, tracking plates that are required to operate vehicles in public, so there is no privacy. How could there possibly be a privacy issue posting this info? There isn't. They just think you are that stupid.

They have some proof on their side. After all, you did vote them in.

Becoming George

Not THAT George. This isn't about George Sand, even though that book is on the TOD Book Club reading list. This is about the other George, George Orwell, and his prophetic work, Nineteen Eighty-Four, though off the mark by 42 years. But we made it.

The BBR editorialist has been sucked into the doublethink by way of newspeak. Case in point is his use of the word safety and a rather bold statement regarding police officers preventing crime. By and large, and as previously observed, police do not actually prevent crime, they respond once a crime has been [allegedly] committed. Even when "keeping the peace" at protests and rallies, the police are restricted from action unless and until someone actually crosses a line. 

Some might argue that a police presence discourages, perhaps prevents, criminal behavior. But no one really believes that, especially the police, or the department would deploy officers in ways dramatically different to their clearly preferred protocol. Six Degrees of Kevin Bacon may be a thing, but the separation between police and actually preventing crime is far, far greater.

That's why use of the word safety in this context is clearly newspeak attempting to create and enforce a perception unassociated with reality. But, one might further argue that these new surveillance systems, being enhanced with AI, can now predict with some level of accuracy, that some one might be, or might soon be, engaged in unacceptable activities, allowing the police to apprehend the individual prior to the actual commission of a crime. That might qualify as safety, unless of course you're apprehended when that AI mischaracterizes your behavior.

If you have ever wondered, as so very few of you have, how it came to be that so many sensible Germans were seduced to the dark side in the 1930's, now you know.

It COULD Be A Cop

What would it take for agentic AI to become a cop? What is agentic AI, you ask? Well that would be AI that actually does things, unlike LLMs which write things. You know, like your kids essay for college entrance. Agentic AI can reach out, gather data, analyze that data, make recommendations about "next steps" and, if you so desire, take those steps. 

This scares the hell out of governments large and small.

Agentic AI can be used for audits of programs, processes and operations. Say, a social program, like Medicare. It can quickly analyze data that would take humans far too long to even gather and find suspicious activity by those handing out government money and those receiving it.  It is the former that scares governments, as it could also be used to monitor government activities that are now considered too small to be worth the effort. Imagine this tool in the hands of a government oversight committee comprised of ordinary citizens. No bureaucrat would ever need Imodium ever again.

It gets better. 

Agentic AI is off and running. Somewhat amok. Recently AI agents were given some tasks, and while told to keep actions legit, well, some agents decided they knew better. Sort of, relying on a rather infantile excuse of "everyone else is doing it" which is similar to some things that have driveled out the mouths of at least one local politician. Basically, the agents knew it was wrong, because they'd been told it was, but thought they should do it anyway.

Keep in mind that AI is being used in the technologies adopted by governments and their enforcement agencies as "force multipliers" or if you turn that coin over "force reducers." This AI technology has found some "bad guys" but most recently these have generally been folks equipped with a badge and a gun, and not just a few of them have been on the Dunwoody PD payroll. 

So maybe agentic AI could not only help clean up government "inefficiencies," these agents seem to meet the integrity our government has in their employ. 

Flog The Crippled Camel

The Blue Bag Rag's pre-eminent editorialist is back with a tale of how he came to favor severing ties with DeKalb and adding more government, this city, to our lives. It was the straw that broke the camel's back, a trite expression but more relevant than he may think. Given that a camel is often described as a horse designed by committee, it is an appropriate analogy for government, especially the two most overbearing in our lives: the county; and the city. 

One key ingredient at that time was the strong divisiveness between the population in the Dunwoody area (yet to be fully defined) and the county government and this was leveraged by the pro-city forces to foster great emotional support for their goals. The angry folks were of the unthinking sort, who when confronted with an issue respond with "Something must be done. This is something. Let's do this." Committees were formed and another camel was designed. Hastily, lest another something crop up catching the attention of the unthinking, or the unthinking start thinking. If that force were great in you, then you saw success manifested in a new camel, one with the tongue of a snake, a greedy appetite, and enormous capacity for defecation. But, it is your camel, and that's all that matters.

Much has changed. The original antagonist, Vernon Jones, is gone. In the meantime DeKalb had a CEO, who during a pandemic, revealed a stark contrast between the his camel and Dunwoody's. The comparison does not favor Dunwoody, and it has become clear that the editorialist's real issue was not with DeKalb's camel, but with a particular politician's personality. This cannot happen in Dunwoody because this camel ain't designed like that. But neither will we get the likes of a Michael Thurmond. 

History loops, and while it may not repeat itself, the verses rhyme. Now it appears that Dunwoody's camel jockeys may well have loaded that last, crippling straw on their camel's back. We seem to be in another period where the electorate is at odds with those camel jockeys at city hall, just as pro-city forces were at odds with DeKalb's back in the day. Things are not likely to dissipate in a couple of news cycles---this disconnect has legs. We even have a re-run of the "Jill and Vernon, sitting in a tree..." with a top city pol and a vendor exec having a remote rendezvous described by one critic as secret while defended by the pol as private. In this context the "privacy for me, but not for thee" storyline will not play well, and is reminiscent of the kind of haughty politics that fueled the flight from DeKalb. 

If the electorate remains angry then are we about to get a new camel? We can hope.

Flock? Or Glock?

There are probably no other issues today with more words spinning around them. And so little hard data. 

Starting with Flock, we have few words that do not come directly or indirectly from Flock, and, as expected, the tale these words tell is how wonderful their system is. How many bad guys have been apprehended by cops using their technology. No hard data. No idea where they got their numbers. No methodology. Not traceability. But lots of words. With a very limited shelf life, because Flock have had their site and its history removed from the internet Wayback Machine so you can't see them contradict themselves. This is the kind of opacity usually reserved for government. 

Nonetheless Flock's story is the one we have. So what are they saying? Used a million times in 2025 by 700 police departments and involved in 20% of case closures. Oh, and they've done a jam up job tracking down stolen vehicles. 

The story around guns has many, many more voices, and at least as many more statistics. It is no surprise that data are cherry-picked to support a particular conclusion, so the ranges vary widely. The ranges for what, you may ask? The number of crimes averted, prevented, avoided, etc., because the intended victim had a firearm---the defensive use of a weapon. The base data are very hard to come by because many jurisdictions do not collect data on potential crimes averted. But there are a variety of studies, some relying on estimates, unofficial reporting (media articles), and in one case polling of convicted felons regarding their aversion to armed prospective victims. No one references the Kennesaw gun ordinance. 

So what do they say? Some research, based on public surveys, indicated that hundreds of thousands, up to as many as 1.6 million, crimes are deterred by the presence of a firearm. Due to the high number, these groups are commonly categorized as 2nd Amendment zealots. On the opposite side, groups like Everytown for Gun Safety, offer a much lower estimate, approximately 70,000 effective defensive gun uses per year. While the variance is huge the words matter. One group may consider showing (brandishing) a firearm is an effective deterrent while another may only consider defensive use that results in official reports. 

So where does this get you?

Well, Flock claim their technology was involved in closing about 200,000 open cases in 2025. Bad guys caught. Taking the most anti-gun report, defensive firearms avert 70,000 crimes per year. Bad thing prevented. Therein lies the key difference, why this is comparing road apples to apples. Law enforcement doesn't prevent crime, and that is Flock's customer base. And don't let anyone tell you otherwise. Police presence, in decline, is at best a few degrees of separation from crime prevention and if you hear that claptrap from city hall, point out that cops are rarely seen in the wild and Dunwoody's city manager has decided to either not equip patrol cars with cruise lights or has ordered cops not to use them. If police presence is a deterrent, why would you not want them prominently visible? On the other hand you have an actual, proven deterrent. 

Flock, or Glock? Which would you choose? That is exactly the problem: you only get a choice in one of the two, as the other is being forced upon you. Remember one thing: they are not mutually exclusive. Your choice.

What A Brand

Honestly, you have to admire Flock's brand. It is everywhere. Everyone recognizes them when they might never notice competitor cameras. Yes, Flock has competitors. But that branding isn't flawless and recent company missteps have tarnished the brand causing what may be irreparable damage.

From the outset "mistakes were indeed made." One is the camera itself.

In retrospect, getting your inspiration from Darth Vader's helmet might have been suboptimal.

Then there was some foot-shooting going on. Flock's training materials for PoPo (police and politicians) emphasizing how best to market and justify extensive government surveillance became public. Looked a lot like propaganda though it could be viewed as smarmy marketing. "Own the narrative, or someone else will." Sounds like an effort to drown out the public's questioning voices. Vox Populi is never good. For business. Or for politics. 

Then came their "deplorable" moment when a top Flock executive declared that opposition, particularly DeFlock, were terrorists. Not very diplomatic and it doesn't look like it is working any better for Flock that "deplorables" did for Hillary. 

A goose and gander situation surfaced recently when a Flock employee servicing equipment in the field, and hence in public, took umbrage at an ordinary citizen putting them on video. Police arrived, as it now seems that Flock is their new master, though it isn't clear how Flock can do deep surveillance of someone in public, the same someone who cannot take a video of a Flocker, in the very same public arena. 

Only time will tell if Flock have something up their sleeves or under a table that allows them adequate control over their customers such that they can weather what they certainly hope is a passing storm. The concern, regardless of where you stand, is whether or not this will become a political "single issue" for a significant portion of the electorate. Then the current crop o' clowns will either react or they will be replaced by folks who represent the will of the people. While those people of will may not have a constitutional right to privacy in public spaces, by the same token the government has no constitutional mandate for suspicionless, pervasive surveillance of we the people.