By: Jennifer Gilligan, IntegraMSP President
Revolutionary technology has always come with risks. How we manage them matters.
During a conversation last week, a peer raised an interesting comparison about AI and how it harkens back to another technology that completely changed society..
Cars.
We depend on them to get to work, connect communities, transport goods, and create opportunities. They've fundamentally changed how we live. They're also extraordinarily dangerous.
According to the National Highway Traffic Safety Administration (NHTSA), an estimated 36,640 people died in traffic crashes in the United States in 2025. Globally, the World Health Organization estimates approximately 1.16 million road traffic deaths annually. Yet we aren't seriously considering abandoning automobiles. We recognize their value, even while acknowledging the harm they can cause.
Over time, we've learned something important: Accepting that a technology has risks isn't the same as accepting that those risks should go unmanaged.
There's a valuable lesson here about artificial intelligence.
Progress Has Never Been Painless
We tend to remember transformative technologies for what they eventually made possible, not necessarily what happened along the way. The automobile expanded economic opportunity, personal freedom, and access to services. It also displaced industries, reshaped communities, introduced environmental problems, and created new dangers.
The benefits were extraordinary. So were some of the consequences. Both things can be true.
Artificial intelligence presents a similar tension. Its potential to accelerate research, improve healthcare, increase accessibility, and help businesses operate more effectively is significant. So are the concerns about privacy, misinformation, bias, cybersecurity, and employment.
And we shouldn't minimize the human impact.
When someone's livelihood is disrupted by automation, telling them that technological progress is inevitable offers little comfort. Innovation creates opportunities, but those opportunities and their associated costs aren't always distributed equally. Recognizing that reality doesn't make someone anti-technology. It makes them thoughtful about its consequences. The existence of risk doesn't automatically negate a technology's value. But neither does that value excuse the harm it may cause.
We Didn't Stop Driving. We Learned to Drive More Safely.
Automobiles didn't become safer because we ignored their dangers. They became safer through engineering, research, regulations, driver education, and accountability. Seat belts, airbags, crash testing, road improvements, and safety standards evolved as we better understood what could go wrong. Some improvements came through innovation. Others followed preventable tragedies and public pressure.
We are still improving. And responsibility doesn't fall solely on the person behind the wheel. Manufacturers, regulators, engineers, and infrastructure planners all play a role in making transportation safer.
AI deserves the same approach to shared responsibility.
We expect people to understand how to operate a vehicle. Shouldn't we expect employees to understand the capabilities and limitations of AI tools?
We expect manufacturers to test safety systems. Shouldn't technology providers evaluate their products for foreseeable failures and harmful outcomes? We establish rules governing where and how vehicles operate. Shouldn't businesses establish boundaries around what information AI can access and which decisions it can influence? And when something goes wrong, shouldn't we know who is responsible? These aren't obstacles to innovation. They're part of making innovation sustainable.
Of course, AI and automobiles are fundamentally different technologies. AI can affect millions of people almost instantaneously, and its consequences aren't always visible or easily reversed. That makes thoughtful oversight more important, not less.
The Cost of Standing Still
There's another side to this discussion that often gets overlooked. Risk doesn't exist only in adopting new technology. There can also be consequences to refusing to explore it. An organization that dismisses AI entirely may overlook opportunities to improve service, reduce repetitive work, identify problems earlier, or help employees focus on more meaningful responsibilities.
For smaller businesses, AI may provide access to capabilities previously available only to much larger organizations. For healthcare professionals and researchers, it can help identify patterns that might otherwise be difficult to recognize. For people with disabilities, it can remove barriers to communication and information. Those benefits matter, too.
None of this means every AI application is worthwhile. Some aren't ready. Some aren't appropriate. Others may introduce risks that outweigh their potential value. But refusing to evaluate a technology is also a decision, with consequences of its own. The question shouldn't be whether AI is universally good or bad. It should be whether a specific application creates meaningful value, what could go wrong, who might be affected, and whether those consequences can be responsibly managed.
Sometimes the right answer will be no. Responsible leadership needs to be comfortable with that answer.
This Is a Leadership Problem as Much as a Technology Problem
In my work, I see organizations focus heavily on what a new technology can accomplish before considering what needs to be in place to support it. AI has made that tendency particularly obvious. An employee discovers a useful tool. A department begins automating a process. Someone connects an application to company data. Before long, AI is influencing business operations, but nobody has established clear expectations around privacy, accuracy, oversight, or accountability.
The technology may be revolutionary. The management challenge isn't.
We've always needed to understand our tools, train the people using them, establish reasonable safeguards, and take responsibility for the outcomes. AI simply introduces new complexity and raises the stakes. And responsibility must extend beyond individual users.
Technology providers have obligations to design and test their products responsibly. Businesses must establish appropriate controls and provide training. Leaders must consider not just efficiency and profitability, but the people affected by their decisions. That includes employees whose roles may change. If an organization benefits financially from AI-driven efficiencies, it should also be willing to consider how it prepares its workforce for those changes.
Not every disruption can be prevented. But treating people as an afterthought shouldn't be an acceptable cost of progress.
For business leaders, responsible adoption starts with some fairly straightforward questions:
- What problem are we actually trying to solve?
- What meaningful benefit do we expect?
- What could go wrong, and who would be affected?
- What safeguards and human oversight are necessary?
- Who is accountable when the technology fails, or circumstances change?
These are practical considerations, not theoretical ones. The National Institute of Standards and Technology's AI Risk Management Framework provides a structured approach to assessing and managing many of these concerns.
What We Can Learn From the Automobile
The comparison isn't perfect, but that's not really the point.
Cars transformed society, and we're still dealing with the consequences more than a century later. We've made them safer through better engineering, clearer expectations, and greater accountability. We've also learned that some problems can't be solved by technology alone. AI will require similar work, although its speed, reach, and complexity present different challenges.
The challenge for businesses isn't simply whether to adopt AI. It's understanding where it adds value, where it introduces risk, and what responsibilities come with using it. We don't need to pretend that AI is harmless to recognize its potential. And we shouldn't dismiss its potential because we haven't solved every problem associated with it.
We've learned to make those distinctions with other technologies. Now we need to apply that experience to AI.
Sources and Further Reading
1. National Highway Traffic Safety Administration (NHTSA)
2025 Traffic Death Estimates & 2024 FARS, April 1, 2026.
https://www.nhtsa.gov/press-releases/traffic-deaths-2025-early-estimates-2024-annual
2. World Health Organization (WHO)
Road Traffic Injuries, updated July 20, 2026.
https://www.who.int/news-room/fact-sheets/detail/road-traffic-injuries
3. Centers for Disease Control and Prevention (CDC)
Global Road Safety, August 6, 2025.
https://www.cdc.gov/transportation-safety/global/index.html
4. National Institute of Standards and Technology (NIST)
Artificial Intelligence Risk Management Framework (AI RMF 1.0).
