The failure to recognize correlations between different risks can lead to a significant underestimation of the total risk facing an organization. In reality, risks do not exist in isolation and there are often interactions between different types of risks. For example, there can be a correlation between market risk and credit risk. If these correlations are not taken into account during the aggregation of risks, it can result in a gross underestimation of the total risk. Therefore, recognizing and accounting for these correlations is crucial for accurate risk aggregation.

Choice A is incorrect. The presence of too many autonomous business units does not necessarily lead to inaccuracies in the calculation of aggregate risk. While it may make the process more complex, as long as each unit accurately assesses and reports its risks, the overall risk can be correctly aggregated.

Choice B is incorrect. The use of different risk assessment models by different business units can potentially lead to inconsistencies in risk measurement across the organization. However, this does not inherently result in inaccuracies in calculating aggregate risk if these models are properly calibrated and validated.

Choice C is incorrect. Recognizing benefits of diversification across an organization actually improves accuracy in calculating aggregate risk because it takes into account that some risks may offset each other due to their negative correlation.