Welders

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

• Carbon arc1881–1927

  Carbon arc and bare metal electrode (experimental era)

  The first arc welding processes (de Méritens carbon arc [1881], Benardos-Olszewski carbon arc of metal [1885-87], Slavyanov metal electrode [1888]) produced results that were technically welds but practically unreliable. Without atmospheric shielding, the molten pool absorbed oxygen and nitrogen from the air, making the weld brittle and porous. The bare-wire deposit was unpredictable under production conditions. The early welders who used these processes were experimenters and repairers rather than production workers; the skill was in managing an unstable process well enough to produce a joint that held. The Kjellberg coated electrode (ESAB, founded 1904) pointed the way forward: wrap the wire in mineral compounds that would generate a protective gas shield and slag layer as they burned. By 1912 Strohmenger had a heavily coated electrode in commercial production, though at too high a cost for broad adoption.

  Effect on the work

  This era established arc welding as a recognizable craft but not yet an industrial one. The welding workforce was small, probably tens of thousands in the US by WWI, concentrated in shipyards and heavy industry repair shops where riveted joints were being supplemented or replaced.

• SMAW shielded metal arc welding1927–1948

  SMAW shielded metal arc welding: extruded electrode coat (1927)

  The 1927 development of an extrusion process for electrode coating changed everything. For the first time, electrode manufacturers could produce complex multi-compound mineral coatings at industrial scale and controlled cost. The coated electrode (a steel core wire wrapped in a precisely formulated blend of silicates, carbonates, and alloying agents) generated its own shielding gas and slag layer as it burned, protecting the molten pool from the atmosphere. The resulting SMAW process (stick welding) was fast, portable, and trainable. It required no external gas supply, just a power source, an electrode, and a trained pair of hands. By the late 1920s, Liberty Ship design was being adapted to accommodate all-welded hulls rather than riveted ones; by WWII, welding had replaced riveting as the primary hull-joining method in US shipyards. The workforce expanded explosively to meet war demand, including the crash-trained welders, "Wendy the Welder", who filled shipyard production lines.

  Effect on the work

  The extruded electrode enabled SMAW to become an industrial production method rather than a repair technique.

• GTAW1941–1960

  GTAW (TIG, 1941) and GMAW (MIG, 1948): inert-gas shielded processes

  Two new welding processes introduced in the 1940s opened the profession to materials and applications that SMAW could not reliably handle. Gas tungsten arc welding (GTAW, or TIG) was patented by Russell Meredith at Northrop Aircraft in 1941 for welding the magnesium airframe of the experimental XP-56 aircraft. TIG uses a non-consumable tungsten electrode and an inert gas shield to protect the weld pool; the filler rod is fed separately by the welder's free hand. TIG produces exceptional weld quality on aluminum, stainless steel, and titanium, the materials of aerospace, and required a level of dexterity and concentration that made it the prestige process of the craft. Gas metal arc welding (GMAW, or MIG) was developed by the Battelle Memorial Institute in 1948 using a continuously fed wire electrode; the 1953 adaptation to carbon dioxide shielding gas made MIG economically viable for steel welding. MIG is faster than SMAW, produces less smoke, and is significantly easier to learn. For auto body sheet metal, MIG became the dominant process by the 1960s, and, critically, the process that robotic welding cells would use when they arrived a decade later.

  Effect on the work

  TIG and MIG expanded the materials welders could work with and opened aerospace manufacturing to the welding trade. TIG created a skill tier that commanded wage premiums. MIG, by being teachable and fast, also enabled the eventual robotic substitution it would power.

• Unimate at GM1961–1985

  Unimate at GM (1961): industrial welding robots begin (spot welding only)

  On June 13, 1961, a 4,000-pound hydraulic arm called Unimate was installed on the assembly line at General Motors' Inland Fisher Guide Plant in Ewing Township, New Jersey. Designed by George Devol and commercialized by Joseph Engelberger through Unimation, the Unimate performed spot welding of auto body parts and transfer of die castings. Spot welding, pressing two overlapping steel panels between copper electrodes and passing a high current through them until they fuse, is geometrically simple and highly repetitive, making it ideal for early teach-and-playback robot programming. For the first two decades, robotic welding was confined to this corner of the craft. Continuous arc welding of structural members, pipe, or complex geometry remained entirely manual. The Fanuc-GM joint venture (Fanuc Robotics) was formed in 1982, accelerating robot adoption in US auto plants through the 1980s.

  Effect on the work

  Robotic spot welding began displacing automotive assembly welders in the 1970s and 1980s, but only in the spot-weld-on-an-assembly-line segment. Structural welders, pipe welders, and job-shop fabricators were entirely unaffected. The net effect on total welding employment was modest through 1985; the displacement was concentrated in a specific sector of a large and varied workforce.

• Automotive spot welding ~95% robotic1985–2000

  Automotive spot welding ~95% robotic: Fanuc, Yaskawa Motoman, ABB dominate

  Through the 1980s and 1990s, robotic spot welding went from a novelty to the dominant method in automotive body assembly. The robots came primarily from three suppliers: Fanuc (Japan, with GM relationship), Yaskawa Motoman (Japan), and ABB (Sweden/Switzerland), later joined by KUKA (Germany). Simultaneously, robot arc welding, continuous bead welding, began growing as programming and sensing improved. By 2005, more than 120,000 robots were in use in North American industry, approximately half of them for welding. The human welding workforce remained large because structural, field, pipe, and custom fabrication work was still entirely manual.

  Effect on the work

  The automotive body-welding workforce was largely displaced from assembly plants in this era, but many welders found work in adjacent applications (truck and heavy equipment manufacturing, structural steel fabrication, bridge construction) that had not been automated.

• Digital inverter welders2000–2015

  Digital inverter welders (Lincoln PowerWave, Miller) + cobots enter fab shops

  The digital inverter power supply (replacing the heavy transformer-based welders of the mid-20th century with compact, precise electronics) transformed the welder's primary tool. Lincoln Electric's PowerWave series and Miller Electric's inverter welders introduced microprocessor control of output waveform, enabling process modes (pulse MIG, AC TIG, RMD) that had previously required specialized equipment. A single machine could now handle multiple processes and materials with parameter recall. Simultaneously, collaborative robots (cobots) began entering small-to-mid-size fabrication shops: lighter, lower-cost systems deployable without traditional robot safety fencing. Lincoln Electric and Miller welding cobot systems brought robotic arc welding within reach of job shops. Cobot welding expanded the automation frontier into structural and fabrication work, though complex joint geometries, tack-fitting, and field conditions still required human welders.

  Effect on the work

  Inverter welders improved portable field-welding capability. Cobots began penetrating job-shop fabrication but remained limited to well-fixtured, repeating joints. The segment of welding work that was automatable continued to grow; the segment that was not (field, repair, complex geometry, short-run) held steady.

• AI arc tracking + orbital pipe robots +…2015–present

  AI arc tracking + orbital pipe robots + IIJA infrastructure demand (2021)

  The most recent era brought incremental automation into previously robot-resistant segments: orbital welding machines for pipe (a motorized head that travels around the pipe's circumference while the welder monitors parameters), adaptive arc-tracking systems that use vision or electrical feedback to follow irregular joint geometries, and AI-assisted parameter optimization. Yet the fundamental manual-weld segments remained intact. The Bipartisan Infrastructure Law (IIJA, 2021) committed $27.5 billion through the Bridge Formula Program specifically, and bridge welding, structural steel or reinforcing bar, remains a manual process. The AWS Certified Welding Inspector program (founded 1976, 100,000+ CWIs certified) certifies human inspectors, not machines. The 2023-2024 welder shortage of approximately 80,000 documented by AWS reflected this structural demand: the automatable corner of welding was automated long ago; the remaining workforce is needed precisely where robots cannot reach.

  Effect on the work

  BLS projects employment growth 2024-2034: modest but positive, defying earlier projections of decline. The welder shortage (approximately 80,000 unfilled positions in 2023-2024) and sustained IIJA infrastructure demand suggest the occupation has stabilized at a floor defined by field welding, pipeline welding, repair welding, custom fabrication, and shipbuilding.