I think one challenge with AI discussions is that the topic is so broad that conversations can become difficult to ground operationally.
“Can anything go wrong with AI?” can span everything from:
hallucinations and decision quality
governance and compliance
data leakage and security
organizational dependency
change management
automation bias
or unrealistic implementation expectations
In practice, I’ve found the most valuable discussions happen when the question becomes more specific and tied to operational realities.
For example:
Where have you seen AI create unexpected delivery or governance risks?
What controls are necessary before AI-generated outputs can be trusted operationally?
How should organizations validate AI recommendations in high-stakes environments?
What signals indicate an organization is adopting AI faster than it can absorb culturally or operationally?
Those tend to generate more actionable insights because they connect AI to real implementation and decision-making challenges rather than treating it as a purely abstract concept. Saving Changes...
Program Manager| HARPER SRLSanto Domingo / Distrito Nacional, Dominican Republic
Hello Saleh Soulyman Yes, many things can go wrong with AI if it is used without enough validation, oversight, or context. Some common risks are:
Incorrect or hallucinated information
Bias in decisions or recommendations
Security and privacy issues
Overdependence on AI outputs
Lack of accountability when mistakes happen
Poor decisions caused by low-quality data
Ethical concerns around transparency or fairness
In organizations, one of the biggest risks is people assuming AI is “always right” because it sounds confident. That’s why human validation and clear governance are still very important. Saving Changes...
Senior Projects Manager | Field & Marten AssociatesNew Westminster, British Columbia, Canada
Saleh, yes, like any powerful tool, AI can go wrong when people rely on it blindly without checking results. AI can make mistakes, reflect biases, generate false information confidently, or be misused in harmful ways. The biggest risk is not necessarily the technology itself, but people placing full trust in its outputs without human judgment, validation, and accountability. AI works best as a tool to assist decision-making, not replace critical thinking. Saving Changes...
Yes, AI can go wrong if it provides inaccurate data, biased recommendations, or makes decisions without proper human oversight. For example, an AI scheduling tool may incorrectly prioritize tasks or allocate resources, causing project delays, budget overruns, or stakeholder conflicts. Saving Changes...
Community Champion
