Physics professors

Three parts of your work where AI is already doing real lifting, and what stays yours.

• Lecture1727–1884

  Lecture, demonstration apparatus, and chalkboard (natural-philosophy era)

  For more than a century of the role, the physics professor's primary tools were the same as any other academic: spoken lecture, handwritten notes, and a stock of physical demonstration apparatus. The natural-philosophy professor of a colonial college kept barometers, air pumps, electrical machines (Leyden jars), and orreries to make visible what the lecture described. Benjamin Franklin's electrical experiments in the 1750s were literally a professor's demonstration at a higher level of ambition. The chalkboard entered American classrooms starting around 1800 and became the physics lecture's dominant medium for equations and diagrams. No amplification, no replication: if a student could not attend the lecture, the knowledge was lost to them.

• Dedicated physics laboratory1884–1940

  Dedicated physics laboratory (Jefferson Lab 1884; Cavendish 1874 model replicated across US)

  The opening of Harvard's Jefferson Physical Laboratory in 1884 -- the first US university building designed specifically for physics research and instruction -- codified the laboratory as essential infrastructure for a physics department. Over the next four decades, every major research university built or expanded a physics laboratory. The laboratory gave the physics professor a new defining task: supervising hands-on experimental work. Students now encountered physics not only through lecture but through measurement, calibration, error analysis, and the irreducible gap between a theory's prediction and a real instrument's reading. This pedagogical model -- teach the theory in lecture, test it in lab -- structured physics education for more than a century and remains essentially intact in 2026.

  Effect on the work

  The laboratory model created demand for a new type of support worker -- the physics laboratory instructor or demonstrator -- and established the need for multiple faculty specializations in what had previously been a one-person natural-philosophy post. Research universities began hiring experimental physicists and theoretical physicists separately during this era.

• Post-Manhattan Project instruments1945–1975

  Post-Manhattan Project instruments: oscilloscopes, vacuum-tube electronics, nuclear detectors (Cold War era)

  The Manhattan Project transformed physics from an academic discipline into a strategic national asset. After 1945, surplus war-era instruments flooded university laboratories: oscilloscopes, signal generators, Geiger-Mueller tubes, and cloud chambers that had been built for the bomb program became undergraduate teaching tools. The National Science Foundation (established 1950) and the Atomic Energy Commission funded new laboratory equipment at a scale universities had never seen. By the late 1950s, Sputnik-era anxiety triggered the National Defense Education Act (1958), which funneled federal money into creating more physics PhDs and university professors. The physics professor of 1960 was using instruments that were simply unavailable in 1930: transistorized oscilloscopes, vacuum-tube amplifiers, and the first digital counters for nuclear experiments.

  Effect on the work

  The AIP Placement Service recorded hundreds of physics jobs for hundreds of new PhDs in 1962 -- an extraordinary seller's market for physics faculty that lasted until 1967-68, when the Vietnam-era de-funding of university research reversed the trend almost overnight.

• Programmable calculator1975–1995

  Programmable calculator and early personal computer (HP-35, TI-58, Apple II, IBM PC)

  The HP-35 scientific calculator launched in 1972 at $395 and rendered the slide rule obsolete within five years. Physics students could now solve transcendental equations and compute numerical integrals at their desks without tables. The Apple II (1977) and IBM PC (1981) brought BASIC programming into undergraduate physics labs: for the first time, students could write simple simulation code to model harmonic oscillators, projectile trajectories, and numerical ODE solutions on machines available in campus computer labs. This era required physics faculty to add programming literacy to their teaching toolkit -- a skill that was not part of physics training before 1970. The symbolic mathematics system Macsyma (MIT, 1968) and later Mathematica (1988) and Matlab (1984) began automating the algebra and calculus that had previously been taught by hand.

  Effect on the work

  The programmable calculator did not reduce physics faculty headcount but did change what a physics problem set looked like: problems involving tedious numerical iteration that were previously impractical as homework became routine. This increased the conceptual difficulty that faculty could demand from students at the same time it lowered the computational barrier.

• World Wide Web1994–2015

  World Wide Web, course management systems, and online problem sets (WebAssign 1997, PhET 2002, Blackboard)

  The World Wide Web reached university campuses in 1993-94 and within a decade had restructured how physics courses operated logistically. WebAssign (North Carolina State, 1997) automated homework submission, grading, and immediate feedback for numerical physics problems -- the first system to partially automate the grading task that consumed significant faculty and TA time. PhET Interactive Simulations (University of Colorado Boulder, 2002) gave students free browser-based interactive physics simulations that ran on any computer, replacing expensive and fragile physical apparatus for introductory demonstrations. Course management systems (Blackboard, Moodle, then Canvas) centralized syllabus, assignments, and grades. By 2010 the administrative overhead of running a physics course had been substantially reduced by technology, even as the instructional core remained human.

  Effect on the work

  Online homework systems reduced TA grading hours significantly in large introductory physics courses and enabled real-time analytics showing faculty which problem types students were consistently missing. PhET simulations reached 250 million uses per year by the early 2020s, demonstrating the scale at which digital tools had supplemented or replaced physical apparatus for introductory physics.

• Gradescope2015–2022

  Gradescope, learning analytics, and AI-assisted grading (Gradescope 2014, Turnitin, Canvas analytics)

  Gradescope (launched at UC Berkeley 2014, acquired by Turnitin 2018) introduced AI-assisted answer grouping for scanned physics exams: the system clusters similar student solutions so a faculty member applies a rubric once per solution strategy rather than per submission, cutting exam-grading time 30-50% on large-cohort courses. These tools did not automate the judgment at the core of physics teaching but substantially reduced the administrative and low-judgment grading labor surrounding it.

• Generative AI physics tools2022–present

  Generative AI physics tools (Wolfram Notebook Assistant, MATLAB Copilot, ChatGPT, Khanmigo)

  Wolfram Notebook Assistant (December 2024) integrated a large language model directly into Mathematica notebooks, enabling natural-language requests for symbolic physics computation across mechanics, electromagnetism, quantum mechanics, thermodynamics, optics, and astrophysics. MATLAB Copilot launched in R2025a (May 2025). ChatGPT and similar models correctly state standard physics principles approximately 95% of the time in controlled tests (arXiv 2411.13685). These tools collectively automate the computational and procedural portions of physics courses that had been the main medium of assessment for more than a century: solving differential equations, deriving expressions from first principles, checking unit analysis, and generating worked examples. Physics faculty who use them save meaningful hours per week on course preparation and solution-key generation; physics students who use them can bypass the procedural practice that was previously the path to physical intuition. The occupation is in the middle of the most significant pedagogical restructuring since the physical laboratory was introduced in 1884.

  Effect on the work

  The 2025 APS/arXiv study (arXiv:2511.11317) documented that physics faculty are actively adapting: using ChatGPT to generate problem-bank variants, redesigning exams around physical interpretation rather than procedural computation, and building custom AI tutors (the Harvard Kestin-Miller model, 2024, doubled learning gains in Physical Sciences 2). The Federal Reserve FEDS Notes (February 2026) places Math/CS at the highest language-model exposure z-score among all college major categories, confirming that the disruption is real and measurable.