Page 9 of Badlands (Nora Kelly #5)

C ORRIE SAT IN her cubicle at the Albuquerque FO, getting her thoughts in order.

With the facial reconstruction complete, she was ready to take the next step in her investigation: letting silicon chips, data farms, and search algorithms help her find the blonde, blue-eyed woman who had—it seemed—left behind her normal life one day, started walking into the desert, and kept on walking until she died.

In recent years, the omnipresence of the internet and the proliferation of databases of personal information had revolutionized the search for missing persons.

Digital footprints were everywhere.

But this case was unusual.

For one thing, by missing person standards, it was old—two to seven years old—and that window was large.

More than half a million people went missing in the US every year.

Those who hadn’t gone the route of Jimmy Hoffa were usually found quickly, but that still left many thousands unaccounted for.

Corrie had a body and a face.

What remained was to link them with an open inquiry.

That meant casting a wide net—a very wide net.

She spent another moment doing her best to put herself in the woman’s head, trying to understand what her thought processes might have been.

She couldn’t have simply vanished without others reporting it, searching, raising hell.

Maybe she was crazy and had committed suicide or been the victim of a crime.

Most likely, she was just lost and delusional from heatstroke.

There’d been no obvious signs of violence.

Nora had been right: the spearpoint had come back negative for any trace residue of blood or human protein—it seemed not to have been involved.

Thinking of the spearpoint led her once again to ponder the two bizarre rocks—lightning stones, Nora had called them—found with the body.

Nora had also said such stones were found only in prehistoric kivas, and that the greenish ones were vanishingly rare.

She’d wondered if perhaps the woman had stolen them and run off, but a quick check showed no such stones were missing, and indeed the only known set today was the one Nora had mentioned.

The woman obviously hadn’t been murdered for them, because they were found by her remains—underneath them, in fact.

The woman had discarded all her clothes, which meant she must have been carrying the lightning stones in her hands.

Bizarre. Beyond the facial reconstruction, those stones were the only important clue they had.

She roused herself. No matter how or why the woman had disappeared, someone must have filed a missing person report.

She just had to find it.

And this digital detection was work she enjoyed: her experience using and misusing computers went as far back as high school, when she’d hacked into the school’s computer system, giving herself straight As and the class bully an assload of Fs.

She preferred doing this kind of hunting at night, when the office was empty and she could slip on her earbuds and listen to Caravan Palace at full volume.

She glanced at the clock.

Maybe she’d end up doing just that—if the search took long enough.

“So: Do we have our man? Or in this case, woman?”

Corrie looked up to see Agent Sharp in the entrance to her cubicle.

“Sir, the game’s afoot.”

“Good.” He came in and took a seat beside her workstation.

“I sent up a second chopper: those badlands cover hundreds of miles, and we want to make sure there isn’t another body out there undiscovered.”

“Thanks.” This was something Corrie couldn’t have authorized on her own.

“And if possible, I’d like to assign Bellamy and O’Hara to the case. I have some thoughts on how they’d be useful.”

“It’s your case.”

Corrie, hearing how he left this sentence hanging, hesitated a moment.

“I was just logging in to NamUs, if… if you’d like to observe.”

His eyes lit up uncharacteristically.

“I’d like that—as long as I won’t be in the way. No doubt you can teach this old dog some new tricks.”

The National Missing and Unidentified Persons System, or NamUs, was a vast repository of data maintained by the Department of Justice to help with cases involving unidentified or missing persons.

Released initially in 200, it was upgraded to 2.

0 in 2018 and again more recently to incorporate AI, remarkably useful for investigations like this.

Corrie had noticed the more senior agents either shied away from it or used it sparingly.

Despite the Bureau’s cutting-edge image, the J.

Edgar Hoover mentality of fingerprint cards and paper files was slow to dissipate, especially among the older agents.

Corrie counted Sharp among those tech-challenged older agents, and casually asking him to watch her at work was, in her mind, a favor for putting her in charge.

He had a quick and curious mind, and she felt sure she could in fact show him some new tricks—and that they’d be appreciated.

As Sharp drew up his chair, she brought up her profile on the central NamUs dashboard, showing draft cases currently part of her workflow—less than a dozen, and none of them active.

Initiating a new search, she entered the woman’s demographic information on a fresh screen—ethnicity, height, gender, last known location.

She was forced to leave some of the most vital data—circumstances, date of last contact—empty, but she filled out a succession of other screens with a physical description and clothing.

More blank fields meant a larger dataset, but she couldn’t help that.

“I’m adding a photo of the facial reconstruction,” she explained.

“It’s not exactly a photo of the person herself, but I can get around that by relaxing the parameters in case I got something wrong in the reconstruction.”

“How large a net are you casting?”

“I thought it would be best to start with the state where the body was found. Depending on results, we can widen the search to the Southwest, or the entire country, if necessary.”

She had now entered all the data available.

Mousing over to the blue search button on the final screen, she clicked it.

After a few seconds, a message came back:

1,304 possible results

“Ouch,” Sharp said.

Just to get a sense of what she was dealing with, she widened the search to include all states in the Southwest.

14,37 possible results

“Angels and ministers of grace defend us,” Sharp murmured.

Fifteen thousand results .

An image of the past came to Corrie’s mind: a vast room smelling of sweat and desiccated paper, library tables arranged in orderly lines, with slanting sunshine and dust motes hanging in the air, a big photo of President Nixon on one wall—and at every table a Caucasian, male FBI agent in a white short-sleeve shirt, poring over fly-specked reports as they all looked for a needle in a haystack.

The FBI had come a long way.

She decided not to bother checking the entire country—not yet, at least. She thought she had an ace up her sleeve that might wipe that laconic expression off Sharp’s face.

Closing the results window, she moved the mouse over to a menu that read ADVANCED TOOLS .

“What’s this?” Sharp asked.

“We’re going to go see the man behind the curtain,” she replied, holding back a smile.

Given her interest in tech, Corrie had been watching the astonishingly rapid rise in artificial intelligence over the last couple of years.

The AI developers were eager to bring their technology from their labs and to the market—while at the same time quelling fears that their AI would become “self-aware” and decide mankind was a parasite to be eliminated…

the usual end to so many sci-fi movies.

To accomplish this, they had gingerly baked AI tools into familiar apps like word processors and spreadsheets, to help with basic tasks like composing letters or creating tables.

But Corrie had been experimenting with release candidates of AI subfunctions included in the NamUs 3.

0 betas. Now was a chance to see what they could do—on a real case.

A new window had opened in the center of the screen—an empty window, containing only a blinking cursor.

On both sides were narrower windows: the search parameters she’d entered on the left, and a hierarchical list of available databases to the right.

“I’ve activated the NamUs copilot,” Corrie said, “and now—since our search is sort of vague—I’m making sure it has all the records and parameters it needs before we unleash it.”

“‘Copilot.’ Oh God—this isn’t Microsoft’s new Clippy, is it?”

Corrie couldn’t help but chuckle at this reference to the software company’s awful attempt, years ago, at creating an on-screen “helper.” “No. In fact, copilot is just a Microsoft marketing term I’m using generically.”

“I’ve heard the proselytizing, but do these AI tools work? How can you be confident they understand what you’re asking?”

“Watch and learn,” Corrie said.

Then—realizing that sounded a little saucy—she quickly added: “If you make sure your search terms are wide enough, and the data pool is large enough, then the AI can do what it does best: encode your request contextually, whittle away the extraneous information and background noise, and build a, um, vector database.”

“I was just going to suggest that,” said Sharp, bemused.

She realized that, in her haste, she was slinging a lot of terms about.

“It’s been trained to search these databases like a human would, making educated guesses, ignoring extraneous information, looking for patterns, and, in a weird way, acting on hunches.”

“A computer can have hunches?”

“Sort of. A human hunch usually comes from combining past experiences to make a best guess. AI does the same thing. It offers up a string of best guesses—except it’s already chewed through more databases than we could in several lifetimes, which gives it ‘experience’ we could never match.”

She began a dialogue with the AI interface:

;:: Are the search parameters I entered properly formatted?

;:: Yes.

;:: Can the LLM process the records without further context?

;:: Yes.

;:: I need a missing person process run, using the given arguments, least degree of accuracy.

Save as a flat file in plaintext format.

The cursor returned to its former blinking state.

“Are you talking to it?” Sharp said.

“Actually, that’s exactly what I’m doing. They call it ‘prompt engineering’—using natural-language comments and corrections to lead the AI to a desired result.”

A response came back on the screen:

;:: Search complete.

712,55 records saved to 0001.

txt

“Aren’t you going in the wrong direction?” Sharp asked.

“I’m just getting it warmed up, so to speak. Now we can start refining the search. The text-to-text model produces increasingly relevant answers each time you instruct it.”

Another brief conversation:

;:: Perform the same run with greatest degree of accuracy

;:: Search complete.

13,442 records saved to 0002.

txt

“Notice how close that result is to the one I got when I tried the normal NamUs search?”

“It’s smaller by 1,500 people—which, I hope, means greater precision.”

“If not, we can always blame technology, right?”

“Work it down to a manageable number, and I’ll upgrade your pool vehicle to a Lincoln Navigator.”

Corrie thought a moment.

“We’ve already given it most of the info we know for sure. That means any of those thirteen thousand, four hundred forty-two records might be our target. There’s no way around that. But we can start adding our own guesses and hunches. Like containing the search to the Southwest, tossing in the possibility that the target was once a model or actress—that sort of thing. Or people who majored in Southwestern archaeology, or have advanced degrees in the subject. They might have knowledge about those lightning stones from studying ancestral Pueblo archaeology.”

“I see where you’re going. How about throwing in Indian Affairs personnel, collectors, and antiquities dealers?”

“Good ideas.” Corrie typed in the filters, in natural language but longer and more specific, listing all the parameters.

She made the AI construct repeat the request in its own words to make sure she’d been understood, then she told it to run the search yet again.

It took over a minute before it responded.

;:: Search complete.

Indexing complete. 1,071 records saved to 0003.

txt

“That’s still a lot of hits,” Corrie said.

“But now, in addition to what we know about the victim, we’ve also put in what we think might be true. But here’s the crucial thing: the software has sorted them by likelihood. So you start with number one and work down, hoping to strike gold long before you reach the bottom of the list. That’s the key difference, what AI brings to the search.”

“Fascinating,” Sharp said.

“And new to me.”

“So,” said Corrie, flushing at the indirect approval, “that’s what Bellamy and O’Hara could be doing—looking through that list.”

Sharp chuckled.

“They’ll love it. May I suggest, for your sake, that I be the one to break the news about their new assignment?” He hesitated.

“As we discussed—red tape, and all that.”

“That would be much appreciated.”

Sharp rose.

“Exceptional, Agent Swanson.”

She hadn’t quite managed to wipe the sleepy look from his face—and, since he’d handed this case to her, the compliment carried less weight than if he’d been actively mentoring her.

But she could tell he’d been deeply impressed—and that was more than enough.