|
Getting your Trinity Audio player ready...
|
The South Coast already knows that training for an unfamiliar piece of equipment begins with safety. Artificial intelligence deserves the same discipline.
In September, Bristol Community College announced construction and skilled-trades safety courses at its National Offshore Wind Institute in New Bedford. That is a practical model for the next wave of workplace technology. A crew must know what could go wrong, how to respond, and who has authority to stop a dangerous process. An AI system that drafts instructions, translates technical material, or summarizes an inspection should face comparable questions.
Many employers want to know how quickly staff can use AI. A better first question is whether the system improves work without making critical mistakes harder to spot.
Gallup’s Oct. 6 findings provide a warning about how unevenly the benefits are arriving. Of American employees who had used AI at work, 63 percent reported completing tasks faster. Across the full workforce, however, just 28 percent used AI at least weekly. AI use was especially limited among production and healthcare-support employees. These national figures do not tell us how any individual New Bedford employer is performing. They do tell us that a rollout built around desk-based workers can leave others out.
That matters in a region with maritime work, healthcare, manufacturing, construction, and hospitality. Consider a supervisor using AI to create a safety briefing. A fluent draft could help, but a missing hazard or incorrect instruction could cause harm. Someone who actually knows the equipment must review the output against the approved procedure. A junior worker should feel free to point out a discrepancy without being treated as an obstacle to innovation.
This month’s AP-NORC survey found that 64 percent of American adults believe AI is developing too quickly. Employers will not earn trust through speeches about inevitable progress. They can earn it by demonstrating what AI may do, what it cannot do, and when a person must take over.
South Coast employers should start with one narrowly defined process. Include the workers who perform it and those who check its quality. Record the baseline error rate, completion time, and number of corrections. Test a draft-generating or information-retrieval tool on approved material, with human verification. Track whether problems become easier to notice and whether inexperienced employees actually learn the underlying task.
Publish the results internally, including failures. Give workers paid time to test improvements. Protect the option to reject an inaccurate output. Avoid using AI efficiency gains as an excuse to eliminate the very training and mentoring that produce competent future supervisors.
New Bedford is building practical training capacity. Employers should build an equally practical standard for AI: safer decisions, stronger skills, and workers with enough authority to challenge a machine.
Gleb Tsipursky, Ph.D., is a behavioral scientist, CEO of Disaster Avoidance Experts, and author of “The Psychology of AI Adoption at Work: From Resistance to Results.”
