Marcus opened his slim black briefcase, withdrew an iPad Pro, and laid it flat on the wood. Without raising his voice, he began issuing directives with the ruthless precision of a surgeon excising a tumor.
"Miss Drummond, you will have the legal compliance department assemble every municipal contract and consulting addendum executed with local police departments over the last five fiscal years in Conference Room B within the hour. Mr. Sterling, I want the complete general ledger for the Public Safety Division, specifically breaking out every promotional stipend, conference sponsorship, and hardware rebate paid to the Ashborne Police Department. And Mrs. Halbrook," Marcus looked at Eleanor, "I want unredacted administrative access to the development architecture and internal correspondence regarding the SafeWatch Civic Intelligence platform."
Robert Kellison, the COO, paled visibly, his tie suddenly looking two sizes too tight. "Mr. Ellison... the SafeWatch algorithmic weights? Sir, that is the crown jewel of our proprietary IP—"
"It is my proprietary IP now, Mr. Kellison," Marcus interrupted evenly. "Fifty-one percent of every patent and every line of code in that vault belongs to Ellison Holdings. And we are going to put on the record right now how that code actually functions."
Kellison swallowed hard, looking around the table for help that wasn't coming. "Sir, SafeWatch is a state-of-the-art predictive neighborhood security ecosystem. It allows verified homeowners and HOA boards to report anomalous activity, and our machine-learning engine cross-references historical patterns to help law enforcement allocate patrol resources efficiently—"
"By automatically flagging Black and Hispanic citizens as high-risk tactical threats based on socioeconomic proxy variables and skin-tone contrast ratios," Marcus finished flatly.
The silence stretched across the boardroom until it became suffocating. Down at the end of the table, CFO David Sterling stared down at his legal pad, his voice emerging barely above a broken whisper.
"The... the engineering team was instructed never to use race as a direct input variable in the code, Mr. Ellison."
"No, you were much cleverer than that," Nora Whitfield said, flipping open a bound forensic report prepared by Marcus’s data scientists. "Instead of a racial input variable, the SafeWatch algorithm weights out-of-zip-code vehicle origin, third-party credit-bureau tiering, facial-recognition non-match profiles at gate cameras, and an automated multiplier called the Verified Resident Credibility Index. Which produces a ninety-four percent identical discriminatory outcome while giving your legal department plausible deniability during municipal procurement hearings."
Marcus tapped the screen of his iPad and swiped a file across the wireless link to the eighty-inch boardroom monitor on the wall. Instantly, a series of scatter plots and dispatch audit tables illuminated the room.
"Before I injected forty-seven million dollars into this company," Marcus said calmly, watching their faces as the data appeared, "my independent audit team obtained eighteen months of anonymized SafeWatch dispatch logs through public-records requests across three municipal police departments currently using your software. In those eighteen months, the SafeWatch algorithm automatically elevated two hundred and fourteen routine civilian sightings to Priority-One 'High-Risk Anomaly' tactical dispatches. Ninety-three percent of the individuals flagged by your software were Black or Hispanic. Eighty-seven percent occurred inside wealthy, majority-white subdivisions where the individuals were legally working, visiting, or—as of this week—living."
Across the table, trapped between her parents' horror and the glowing data on the wall, Vanessa’s lifelong conditioning finally overrode her paralysis. Her chin jerked up, a desperate, defiant flush returning to her cheeks as she gripped the edge of the mahogany table.
"That isn't racial discrimination!" she snapped, her voice shrill and brittle in the vast room. "That's just objective statistical data! The algorithm learns from historical incident baselines! If certain demographics keep showing up in high-value residential sectors where they statistically don't belong—"
"Like me standing in my own two-million-dollar driveway, Miss Halbrook?" Marcus asked quietly.
The question hung in the air like a guillotine blade suspended by a single thread. Vanessa’s mouth snapped shut, her chest rising and falling rapidly as her father closed his eyes in sheer agony.
Marcus tapped his iPad again, bringing up a specific server log stamped with the Halbrook Dynamics seal. "Your division sold this reporting infrastructure to seventeen police agencies in this region, including the Ashborne Police Department eight months ago. Yesterday morning at 8:58 a.m., when Miss Halbrook dialed 911 from 15 Magnolia Drive to falsely report a Black man stealing a Ferrari across the street, the SafeWatch bridge didn't treat her call like an unverified tip. Because her personal cell phone number is registered in the SafeWatch database with an Executive VIP credibility override, the algorithm automatically bypassed standard verification protocols and dispatched Unit 412 under a high-risk felony detention code."
Arthur Halbrook slammed his palm down onto the back of the leather chair, his face purple. "Dammit, Ellison, you're twisting a public safety tool that protects tax-paying communities into some kind of political crusade!"
"I am describing a defective, profit-driven liability engine that monetizes suburban bigotry and turns neighborly paranoia into armed police encounters, Arthur," Marcus replied, his voice never rising a single decibel. "Your daughter didn't just make a racist phone call yesterday morning. She pulled the lever on a corporate machine designed by this company to guarantee that patrol officers would treat a wealthy white woman's lie as more credible than a Black homeowner's property deed."
At the head of the table, Eleanor Halbrook spoke for the first time since Marcus had sat down, her voice tight and brittle as dry ice. "The SafeWatch platform was developed in direct response to security demands from our municipal and residential clients, Mr. Ellison. We provide analytical tools that help understaffed police departments prioritize their field resources—"
"Based on systemic profiling disguised as data science," Nora Whitfield cut in, opening a second red-tabbed binder and reading aloud into the quiet room. "Internal Customer Compliance Log #104, dated March 15th of this year: Maria Santos, a state-certified home health aide, held in handcuffs on the curb for twenty-five minutes while caring for a ninety-year-old stroke patient in Riverside Estates after SafeWatch’s gate camera flagged her ten-year-old Honda Civic as a 'Recurring Low-Tier Anomaly.' Internal Log #119, dated April 2nd: David Jackson, licensed master electrician, detained at gunpoint by two patrol units while unloading conduit at Willow Ridge after a resident app report triggered SafeWatch’s 'Predatory Loitering' protocol."
Nora lifted her eyes from the page and looked squarely at Vanessa. "Shall I keep reading, Miss Halbrook? Our forensic team pulled thirty-seven suppressed civil rights complaints from your department's internal server from the last twelve months alone—every single one of which your PR office buried with non-disclosure settlements or dismissed as 'acceptable false-positive parameters.'"
Vanessa’s hands were trembling visibly now against the polished wood, yet she couldn't stop herself from digging deeper. "Those are isolated statistical edge cases! Any complex security software has false positives! The system works exactly as it was designed to work—"