AI Accountability Failures: What Can Go Wrong | QuizBy Eyal Doron / December 6, 2025 / 1 minute of reading AI Accountability Failures: What Can Go Wrong | Quiz 1 / 7 1. What is the purpose of a RACI matrix in AI accountability? 1. To document who is responsible accountable consulted and informed for each AI decision 2. To assess the technical capabilities of AI team members 3. To calculate the return on investment for AI governance programs 4. To rank AI systems by their risk level for compliance purposes Correct! WHY: A RACI matrix documents who is Responsible Accountable Consulted and Informed for each aspect of the AI system lifecycle preventing ambiguity. CONTEXT: This structured approach unites cross-functional teams under shared responsibility instead of allowing siloed thinking. REMEMBER: RACI prevents the everyone and therefore no one is responsible trap. 2 / 7 2. What was the root cause of the Zillow iBuying collapse according to the article? 1. No one was assigned responsibility for monitoring model drift 2. The AI algorithm was fundamentally flawed from the start 3. COVID market conditions were impossible to predict 4. The company lacked sufficient training data for home prices Correct! WHY: The monitor role was unowned meaning no one was assigned responsibility for watching model drift and escalating concerns to leadership. CONTEXT: This resulted in 569 million dollars in losses and 2000 layoffs because model drift festered for months without anyone accountable for detection. REMEMBER: Models degrade silently and someone must be watching. 3 / 7 3. In the hiring algorithm discrimination case what was the key accountability lesson? 1. AI vendors are solely responsible for bias in their products 2. Deployers are accountable for AI behavior even when the AI learned patterns from data 3. Historical data automatically creates legal protection for AI users 4. Regulatory bodies should audit all AI training data before deployment Correct! WHY: The company remained accountable because humans chose the training data chose to deploy the tool and chose not to audit for bias. CONTEXT: This case demonstrates that deployers cannot transfer responsibility to algorithms by claiming the AI learned from data. REMEMBER: Deployers are accountable for AI behavior regardless of what the AI learned. 4 / 7 4. What quick win does the article recommend for improving AI accountability? 1. Hire a dedicated AI ethics officer for your organization 2. Create an AI ethics committee with monthly meetings 3. Define the accountable owner and required documentation for your two highest-risk AI systems 4. Purchase AI governance software from a major vendor Correct! WHY: Defining accountable owners and documentation requirements for high-risk systems is an immediate actionable step that addresses the most critical accountability gaps. CONTEXT: This focuses resources on the systems where accountability failures would cause the most harm. REMEMBER: Start with your two highest-risk AI systems this month. 5 / 7 5. What is the purpose of assigning a single accountable owner for each AI system? 1. To reduce the number of people who need AI training 2. To ensure someone is unambiguously responsible for outcomes 3. To limit legal liability to one individual 4. To minimize the cost of AI governance programs Correct! WHY: A single accountable owner ensures someone is unambiguously responsible for outcomes even if they do not do all the work themselves. CONTEXT: This prevents the diffused responsibility problem where everyone can point to someone else. REMEMBER: This person may not do all the work but they own the outcomes. 6 / 7 6. What characterizes diffused responsibility as an accountability failure pattern? 1. A single executive taking too much control over AI decisions 2. Technical complexity that prevents anyone from understanding the system 3. Documentation spread across multiple systems making it hard to find 4. Responsibility split across teams so each can blame others when problems occur Correct! WHY: When responsibility splits across multiple teams each team can point to decisions made by others creating gaps where harm falls through organizational cracks. CONTEXT: This is the everyone is responsible therefore no one is responsible trap that paralyzes incident response. REMEMBER: If everyone is responsible nobody is responsible. 7 / 7 7. When a company says the AI did it in response to AI-caused harm what does this represent? 1. A technical explanation for model behavior 2. An accountability failure that deflects responsibility to the algorithm 3. An accurate description of autonomous AI decision-making 4. A valid legal defense in most jurisdictions Correct! WHY: Saying the AI did it is an accountability failure because it deflects responsibility to a tool rather than the humans who chose to deploy it. CONTEXT: Just as we do not say the spreadsheet did it when financial decisions go wrong we cannot blame AI for decisions humans enabled. REMEMBER: AI is a tool and tools cannot be accountable. Your score isThe average score is 0% Restart quiz Download PDF Please leave this field empty🔐 The AI Security Manager's Newsletter Weekly insights on AI risk management, EU AI Act compliance, and practical security strategies. We don’t spam! Read our privacy policy for more info. Thank you! Please check your inbox to confirm your subscription.
