When AI Fails, Accountability Still Belongs to Us

Last week we talked about what happens when everything goes dark. When AWS, CrowdStrike, or another global system collapses, companies rush to explain that it was out of their control. What we found then was a pattern. The more connected we become, the more blame gets outsourced. Now that same story is unfolding again, only this time the failure is coming from the inside. 

The new AI Safety Report from RAIDS AI Limited, shared with me by Nikolas Kairinos, analyzed more than forty real cases where artificial intelligence systems caused measurable harm. The total damage exceeds one hundred billion dollars. These are not theoretical risks or science fiction scenarios. They are real systems already operating in hospitals, financial markets, and customer service, making decisions that shape economies and lives. These examples remind us that progress without accountability is reckless and that AI should extend human judgment, not replace it. 

There was the trading algorithm that lost four hundred forty million dollars in forty-five minutes. The hospital AI that missed infections in patients it was supposed to save. The autonomous vehicle that classified a pedestrian as a shadow and never hit the brakes. The customer service bot that promised refunds its company did not offer, costing both money and credibility. 

In every case, the machine performed exactly as instructed. It did not rebel or make a mistake. It optimized the goal it was given, even when that goal stopped making sense. The failure was not in the code. It was in the leadership. 

This is the new accountability dilemma. When everything is automated, it becomes easier to say the system made the call. But systems do not write strategy, approve budgets, or stand in front of shareholders. People do. 

The report found that AI-related incidents rose by more than fifty percent in the last year alone, yet fewer than one in four organizations actively monitor their AI behavior. That means most companies are operating with blind spots large enough to cause irreversible damage. When an algorithm drifts off course, the problem often goes unnoticed until it has already affected thousands of transactions, diagnoses, or decisions. 

It is the same story we saw with global outages. The more complex and invisible the systems become, the easier it is for leaders to shift responsibility to something abstract. But accountability is not about who flipped the switch. It is about who allowed the conditions for failure to exist. 

The report ends with an aviation comparison. Air travel became safe not through better engines but through a culture of safety. Pilots, engineers, and crews developed checklists, simulations, and communication routines because they learned that technology alone does not create trust. The same principle applies to AI. 

Leaders must build that same kind of cultural infrastructure. Every new system should come with a monitoring plan. Teams should question what success metrics really measure. Curiosity must replace complacency. The technology may be autonomous, but the responsibility never is. 

AI will continue to change how we work, decide, and compete. But when it fails, it will not be because the machine made the wrong call. It will be because we stopped paying attention to the right ones. 

Elsewhere In Culture 

https://fortune.com/2025/08/18/quiet-cracking-workplace-culture-employees-burnout-disengagement-mental-health-billions-business-loss-managers-ai-promotions

Too often, people stay in roles that no longer serve them because they feel obligated, loyal, or afraid to start over. But as Poduška points out, that is exactly how quiet cracking begins, when you have outgrown your environment and your company has stopped investing in your growth. The healthiest thing you can do for your career is to ask hard questions about sustainability and support. Do you feel energized by the work you are doing? Do you have leaders who care about your development? If the answer is no, it may not be burnout, it may be misalignment. 

Quiet cracking does not always require a dramatic exit or reinvention. Sometimes it is about taking control of your own development when your organization will not. That could mean seeking mentorship outside your company, exploring cross functional projects, or finding a leader who recognizes your potential. And yes, sometimes it means walking away. Growth always requires movement, and when the culture around you stands still, that movement may need to come from you. 

https://www.reuters.com/business/world-at-work/amazon-targets-many-30000-corporate-job-cuts-sources-say-2025-10-27

Amazon’s decision to cut 30,000 corporate jobs is being framed as a move toward efficiency, but it also raises questions about stability and resilience. When companies move too quickly to downsize in favor of automation, they often underestimate the cost of losing human judgment. History has shown that early adoption of new technology is rarely seamless. AI systems have already made costly mistakes, from trading algorithms that erased millions of dollars to medical tools that failed to detect infections they were designed to catch. These are not futuristic risks. They are examples of optimization without oversight, and they show what happens when machines follow instructions long after those instructions stop making sense. 

The companies that succeed in this era will not be the ones that automate the fastest but the ones that stay flexible enough to correct course when technology falls short. There is also an economic consequence that goes beyond the company itself. Every time a business replaces workers with automation, it reduces the number of people who can buy its products. Past revolutions created new opportunities, new jobs, and new consumers. The AI revolution, at least so far, is focused on doing more with less. If that pattern continues, the short-term gains that please investors today could lead to a long-term slowdown in demand tomorrow. 

Surrender to Lead is the book Joe Terry and I wrote to redefine what leadership really looks like. We have both seen how often leaders think strength means control, but it doesn’t. Real leadership starts when you let go. When you surrender your ego, your need to have all the answers, and your attachment to certainty, you create space for trust, accountability, and real results. 

Preorders are now live. When you preorder, you’ll also receive three hundred dollars worth of exclusive leadership resources to help you apply these ideas in your own organization. Visit surrendertolead.com to preorder your copy today.