Course Planning: DTPA Use Case Matrix on AI and Society Course

1.1 Automated Essay Scoring (AES)

Thu, 21 May 2026 00:00:00 +0000

Automated Essay Scoring (AES)

Context & Systems Architecture

Automated Essay Scoring (AES) systems utilize natural language processing (NLP) pipelines—ranging from traditional feature-extraction regression models (such as ETS’s e-rater) to fine-tuned transformer networks (such as BERT variants)—to evaluate student writing. These systems are widely deployed by major assessment corporations, including Pearson, ETS, and Turnitin, to grade high-volume standardized tests, admissions essays, and placement exams.

In a significant recent escalation of this technology, the Texas Education Agency deployed an automated scoring engine in 2024 to grade the State of Texas Assessments of Academic Readiness (STAAR) exams. The agency projected that the system would save $15–$20 million annually by reducing the number of human temporary graders from roughly 6,000 to under 2,000. However, the rollout resulted in widespread school district complaints after a dramatic surge in zero-scores for responses that did not fit the model’s expected syntax paths.

1.2 AI Tutoring Systems

Thu, 21 May 2026 00:00:00 +0000

AI Tutoring Systems

Context & Systems Architecture

AI tutoring systems have transitioned from rigid rule-based cognitive tutors to conversational platforms powered by large language models (LLMs). The most visible system is Khan Academy’s Khanmigo, introduced as a pilot in 2023. Khanmigo leverages foundational models (originally GPT-4, updated to GPT-4 Turbo and GPT-4o) bounded by system-level instructions that compel the AI to act as a Socratic guide rather than directly providing answers. The educational goal is to scale individual instruction to mitigate the post-pandemic learning loss.

1.3 AI-Powered Exam Proctoring

Thu, 21 May 2026 00:00:00 +0000

AI-Powered Exam Proctoring

Context & Systems Architecture

Remote exam proctoring applications (such as Proctorio, Honorlock, and Respondus Monitor) surged during the COVID-19 pandemic and continue to monitor high-stakes testing environments. These applications enforce academic integrity by locking down student web browsers and transforming the student’s personal webcam into a computer-vision surveillance apparatus. The software monitors eye movements, head orientation, ambient audio levels, and keyboard rhythms, automatically compiling a timeline of “suspicious events” for instructor review.

1.4 Predictive Student Retention Systems

Thu, 21 May 2026 00:00:00 +0000

Predictive Student Retention / Early Warning Systems

Context & Systems Architecture

Dozens of universities leverage enterprise predictive analytics platforms—such as EAB Navigate360 or Civitas Learning—to optimize student retention rates and allocate advising resources. By integrating directly with institutional Learning Management Systems (LMS) and student information databases, these platforms use historical enrollment logs to calculate a real-time “risk score” for every student, alerting advisors to individuals predicted to fail or drop out.

10.1 Alternative Data Credit Scoring

Thu, 21 May 2026 00:00:00 +0000

Alternative Data Credit Scoring

Context & Systems Architecture

Traditional credit underwriting frameworks (such as FICO scores) rely on structured financial histories, including repayment logs, existing debt levels, and credit card histories. However, across the fintech and subprime lending landscape, algorithmic credit models have emerged that bypass these traditional guardrails. FinTech platforms and neo-banks utilize non-traditional machine learning classifiers to score the creditworthiness of “credit invisible” or underbanked individuals. These systems ingest digital behavior footprints to predict the statistical probability of loan default, deeply complicating standard consumer protection laws.

11.1 Lethal Autonomous Weapons Systems (LAWS)

Thu, 21 May 2026 00:00:00 +0000

Lethal Autonomous Weapons Systems (LAWS)

Context & Systems Architecture

Lethal Autonomous Weapons Systems (LAWS)—frequently referred to in public discourse as “killer robots”—represent a major shift in the execution of state violence. Unlike remote-piloted drones, where a human operator reviews video feeds and pulls a physical trigger, LAWS are robotic systems equipped with onboard sensor arrays and computer vision software designed to select, track, and engage targets with lethal force entirely on their own. These systems encompass everything from loitering munitions (like the STM Kargu-2 or AeroVironment Switchblade arrays) to autonomous marine vessels, operating at speeds that outpace human command-and-control loops.

2.1 Medical Image Diagnosis (Radiology AI)

Thu, 21 May 2026 00:00:00 +0000

Medical Image Diagnosis (Radiology AI)

Context & Systems Architecture

Computer-vision algorithms designed for radiology represent one of the most clinically mature applications of AI. FDA-cleared systems built by companies like Aidoc, Viz.ai, and Google DeepMind analyze chest X-rays, head CT scans, and mammograms directly within hospital workflows. These systems function as parallel diagnostic layers, screening medical imagery to flag acute anomalies—such as intracranial hemorrhages, pulmonary embolisms, or malignant nodules—and prioritizing them on a radiologist’s reading queue.

2.2 Health Insurance Risk Scoring

Thu, 21 May 2026 00:00:00 +0000

Health Insurance Risk Scoring & Care Management

Context & Systems Architecture

Commercial health insurance companies and integrated health networks rely heavily on predictive scoring algorithms to manage large patient populations. These tools generate a “risk score” for each patient to identify individuals with complex, chronic needs for enrollment in high-risk care management programs. These programs grant patients access to dedicated nursing staff, home health visits, and prioritized primary care appointments to prevent sudden hospitalization.

2.3 Mental Health Chatbots

Thu, 21 May 2026 00:00:00 +0000

Mental Health Chatbots

Context & Systems Architecture

Driven by acute clinician shortages and the low scalability of human therapy, conversational mental health applications (such as Woebot, Wysa, and specialized wellness tools) have expanded rapidly. These applications deliver automated Cognitive Behavioral Therapy (CBT) frameworks, mindfulness prompts, and mood-tracking exercises via automated interfaces. While marketed using medical language, these platforms are legally structured as “wellness software” to navigate around rigid healthcare regulations and liability standards.

2.4 Sepsis Prediction & Hospital Triaging Algorithms

Thu, 21 May 2026 00:00:00 +0000

Sepsis Prediction & Hospital Triaging Algorithms

Context & Systems Architecture

Sepsis is a life-threatening medical emergency caused by the body’s extreme response to an infection, accounting for roughly one in three hospital deaths in the United States. To catch signs of deterioration before overt clinical collapse, hundreds of hospitals across the country integrated automated, proprietary predictive models directly into their Electronic Health Record (EHR) ecosystems—most notably the Epic Sepsis Model (ESM) developed by Epic Systems. These tools run continuously in the background of active medical wards, computing an automated probability score of an individual patient developing sepsis and throwing real-time pop-up alerts to bedside nurses and physicians.

3.1 Predictive Policing & Spatial-Temporal Risk Loops

Thu, 21 May 2026 00:00:00 +0000

Predictive Policing & Spatial-Temporal Risk Loops

Context & Systems Architecture

PredPol (later rebranded Geolitica) represents one of the most widely deployed predictive policing applications in the United States, integrated into major metropolitan departments including the LAPD and Chicago PD between 2011 and 2020. Marketed as an objective tool capable of identifying where crimes would occur before they happen, the software promised to optimize municipal resource allocation. However, the system’s foundational architecture obscured a fundamental operational reality: the algorithm does not map actual criminal occurrences, but rather the historical deployment patterns of the law enforcement agency itself.

3.2 Facial Recognition in Law Enforcement

Thu, 21 May 2026 00:00:00 +0000

Facial Recognition in Law Enforcement

Context & Systems Architecture

The integration of computer vision facial recognition tools into law enforcement workflows has fundamentally shifted the nature of police identification. Utilizing both public databases (such as DMV photo repositories and mugshots) and unregulated private scraping systems like Clearview AI, police departments run photos from surveillance clips or mobile devices against millions of identities. While marketed as a pinpoint forensic breakthrough, the real-world execution of these computer vision pipelines has resulted in catastrophic failures, specifically the documented wrongful arrests of innocent individuals due to algorithmic misidentification.

3.3 Risk Assessment at Sentencing (COMPAS)

Thu, 21 May 2026 00:00:00 +0000

Risk Assessment at Sentencing (COMPAS)

Context & Systems Architecture

The Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) is a proprietary machine learning classification tool developed by Northpointe (now Equivant). It is widely integrated into the United States criminal justice apparatus, specifically used by judges, parole officers, and corrections departments in states like Wisconsin, Florida, and New York to guide pre-sentencing reports, bail amounts, and parole determinations. The system is designed to predict a defendant’s risk of recidivism—the statistical likelihood that an individual will commit another crime within a specified window (typically two years)—by computing an automated risk rating.

3.4 Mass Surveillance & Smart City Infrastructure

Thu, 21 May 2026 00:00:00 +0000

Mass Surveillance & Smart City Infrastructure

Context & Systems Architecture

Modern metropolitan centers have increasingly transitioned into “smart cities” by deploying deeply integrated, always-on sensor networks. Facilitated by corporate partnerships with defense and technology contractors like Axon, Palantir, and SoundThinking (formerly ShotSpotter), municipalities have constructed layered surveillance networks. These configurations capture acoustic, visual, and electronic emissions across public urban spaces. While framed as public safety modernization projects, these systems continuously track population movements, fundamentally changing the legal expectations of public privacy.

3.5 Automated License Plate Readers (ALPR)

Thu, 21 May 2026 00:00:00 +0000

Automated License Plate Readers (ALPR)

Context & Systems Architecture

Automated License Plate Readers (ALPR) have quietly transformed municipal roads and highways into a seamless, searchable surveillance dragnet. Built primarily by private corporations like Flock Safety and Vigilant Solutions, ALPR networks utilize optical character recognition (OCR) camera arrays mounted on utility poles, police cruisers, and neighborhood entryways. Rather than tracking individual suspected vehicles under active judicial warrants, these systems capture every single vehicle that passes through their field of view, logging geographical coordinates, precise timestamps, and visual profiles into centralized cloud databases accessible by thousands of law enforcement agencies nationwide.

4.1 Generative Image Models & Artist Rights

Thu, 21 May 2026 00:00:00 +0000

Generative Image Models & Artist Rights

Context & Systems Architecture

The sudden scaling of generative artificial intelligence in visual arts—anchored by commercial systems such as Midjourney, Stable Diffusion, and OpenAI’s DALL-E—relies on deep generative diffusion models. These tools allow users to output high-fidelity illustrations, digital paintings, and photographic assets from simple text prompts. However, the foundational infrastructure of this sector was built on the non-consensual extraction of billions of copyrighted creative works, leading to an intense legal, economic, and ethical conflict between tech conglomerates and the global creative labor force.

4.2 AI and Actor Likenesses (SAG-AFTRA Strike)

Thu, 21 May 2026 00:00:00 +0000

AI and Actor Likenesses (SAG-AFTRA Strike)

Context & Systems Architecture

During the 2023 Hollywood labor disputes, generative artificial intelligence emerged as a major point of conflict, culminating in the historic 118-day SAG-AFTRA actors’ strike. Entertainment studios represented by the Alliance of Motion Picture and Television Producers (AMPTP) rapidly adopted high-fidelity 3D body scanning, generative adversarial networks (GANs), and neural voice cloning. This technology allowed studios to create permanent digital twins of human performers, raising urgent questions about bodily autonomy, digital ownership, and the future of human employment in creative industries.

4.3 AI in Legal Knowledge Work: The Mata v. Avianca Case

Thu, 21 May 2026 00:00:00 +0000

AI in Legal Knowledge Work: The Mata v. Avianca Case

Context & Systems Architecture

The sudden availability of commercial Large Language Models (LLMs) catalyzed intense corporate hype regarding the complete automation of knowledge-work sectors, particularly the legal profession. This narrative faced an unprecedented real-world crisis in the federal case of Mata v. Avianca, Inc. (2023) in the U.S. District Court for the Southern District of New York. Faced with a complex statute of limitations motion to dismiss, plaintiff’s attorneys utilized OpenAI’s ChatGPT to conduct legal research and draft a formal opposition brief. The resulting judicial breakdown exposed the deep mismatch between consumer expectations of AI intelligence and the actual structural mechanics of predictive language generation.

5.1 Climate Modeling & Weather Prediction

Thu, 21 May 2026 00:00:00 +0000

Climate Modeling & Weather Prediction

Context & Systems Architecture

AI-driven atmospheric modeling has emerged as a disruptive paradigm shift threatening to overturn traditional numerical weather prediction (NWP). Historically, weather forecasting relied on massive supercomputers executing complex systems of physical fluid dynamics and thermodynamic differential equations. In 2023, Google DeepMind released GraphCast, a deep learning model capable of generating highly accurate 10-day global weather forecasts in under 60 seconds on a single GPU—matching or exceeding the predictive skill of the European Centre for Medium-Range Weather Forecasts (ECMWF), the historic global gold standard.

5.2 The Carbon Cost of AI

Thu, 21 May 2026 00:00:00 +0000

The Carbon Cost of AI

Context & Systems Architecture

The rapid integration of artificial intelligence into daily digital infrastructure is frequently presented as a clean, virtual shift that saves human labor and carbon footprint. This framing hides a massive, highly material physical reality: AI runs on an incredibly resource-intensive global network of factories, power grids, and cooling infrastructure. The expansion of generative AI and LLM clusters has triggered an unprecedented surge in electricity demand, forcing tech conglomerates to expand fossil-fuel dependencies and directly undermining global carbon reduction mandates.

6-3 Sub-second Productivity Tracking Surveillance

Thu, 21 May 2026 00:00:00 +0000

Sub-second Productivity Tracking Surveillance

Context & Systems Architecture

The traditional manager-worker relationship has been heavily automated in large-scale fulfillment centers and logistics hubs. The pinnacle of this shift is represented by Amazon’s proprietary infrastructure, specifically its automated labor tracking software known historically as ADAPT (Associate Development and Performance Tracker). This architectural framework treats human workers as mechanical units within an algorithmic logistics chain. Handheld barcode scanners, thermal imaging arrays, and smart-vest biometrics track worker physical performance down to the individual second, turning real-time physical movement into continuous performance metrics.

6.1 AI Resume Screening & Hiring Discriminators

Thu, 21 May 2026 00:00:00 +0000

AI Resume Screening & Hiring Discriminators

Context & Systems Architecture

Automated hiring and talent acquisition tools are widely used by corporate Human Resources departments to filter through thousands of applicant resumes. Between 2014 and 2017, Amazon developed a proprietary, secret AI recruiting engine designed to automatically rank job applicants and identify top talent. The project was ultimately abandoned after internal engineers discovered that the machine learning pipeline had developed a systematic, algorithmic hostility toward female candidates, uncovering the core vulnerability of training models on historical human decision-making data.

6.2 Algorithmic Management in the Gig Economy

Thu, 21 May 2026 00:00:00 +0000

Algorithmic Management in the Gig Economy

Context & Systems Architecture

Gig economy platforms like Uber, Lyft, DoorDash, and Instacart have completely replaced traditional human management with algorithmic coordination systems. Workers interact with their employer solely through a mobile application interface that tracks behavior, distributes jobs, determines compensation rates, and administers disciplinary actions. This framework creates an extreme information asymmetry, allowing corporate platforms to exercise total behavioral control over a massive distributed workforce while avoiding the legal and financial obligations associated with employing human staff.

7.1 Recommendation Algorithms & Engagement Optimization

Thu, 21 May 2026 00:00:00 +0000

Recommendation Algorithms & Engagement Optimization

Context & Systems Architecture

Social media platforms like YouTube, TikTok, and Instagram are anchored by hyper-optimized recommendation architectures. These systems utilize machine learning pipelines to curate personalized feeds for hundreds of millions of users. Driven by a commercial model where user attention is directly monetized through advertising impressions, these systems are tuned to optimize for a single target: engagement maximization. The societal side effects of this design choice were exposed in historic whistleblower disclosures, revealing that platforms prioritize corporate growth over user safety.

7.2 Deepfakes & Non-Consensual Synthetic Media

Thu, 21 May 2026 00:00:00 +0000

Deepfakes & Non-Consensual Synthetic Media

Context & Systems Architecture

The democratization of generative artificial intelligence has led to a rapid proliferation of synthetic media, commonly known as deepfakes. Utilizing generative adversarial networks (GANs) and advanced diffusion models, malicious actors can synthesize high-fidelity video and voice clones using consumer-grade computer hardware. While occasionally deployed for creative satire or cinematic visual effects, the primary operational manifestation of this technology has been weaponized as a vector for targeted harassment, political disinformation, and non-consensual sexual violence.

8.1 Algorithmic Rent Pricing (RealPage)

Thu, 21 May 2026 00:00:00 +0000

Algorithmic Rent Pricing (RealPage)

Context & Systems Architecture

Over the past decade, the deployment of automated revenue management systems has fundamentally restructured the economics of urban housing. The most prominent player in this space is RealPage, a real estate software firm that provides dynamic price-optimization software (historically known as YieldStar and upgraded to AI Revenue Management systems).

By the mid-2020s, RealPage’s platform generated daily rental rate recommendations for an estimated 16 million units across the United States, touching a massive cross-section of the domestic multi-family housing market. Rather than relying on traditional localized property manager intuition or public real estate listings, RealPage built a centralized mathematical framework that aggregates proprietary transactional data directly from competing property management corporations to calculate profit-maximizing rental prices.

9.1 Ghost Work: The Hidden Human Labor Powering AI

Thu, 21 May 2026 00:00:00 +0000

Ghost Work: The Hidden Human Labor Powering AI

Context & Systems Architecture

The public marketing narratives surrounding artificial intelligence systems champion the illusion of total technological autonomy, framing tools like ChatGPT as purely automated computational miracles. This framing masks a massive global network of human exploitation. Every supervised machine learning model and safety-filtered LLM is fundamentally dependent on an expansive, underpaid labor force situated primarily in the Global South. This workforce manually reviews, labels, and sanitizes toxic data packets to make artificial intelligence safe and profitable for tech conglomerates.

9.2 Biometric ID Systems (Aadhaar)

Thu, 21 May 2026 00:00:00 +0000

Biometric ID Systems (Aadhaar)

Context & Systems Architecture

India’s Aadhaar system represents the largest national biometric identification infrastructure in human history, encompassing over 1.4 billion registered individuals. Managed by the Unique Identification Authority of India (UIDAI), the system links an individual’s fingerprints, iris scans, and facial photographs to a unique 12-digit identification number. Originally introduced as a modernization initiative to streamline welfare distribution, eliminate administrative corruption, and delete “ghost beneficiaries,” the system has transformed into an mandatory digital gatekeeper for basic survival resources.