The auditor general's report does not find a technology failure — it finds a policy choice. When a government procurement scorecard assigns accuracy a 4% weight and regional presence a 30% weight, the resulting approvals are not a surprise or an oversight; they are the intended output of that scoring regime. Ontario's approved AI scribes that hallucinated treatment suggestions passed a process that was structurally indifferent to whether they worked. The vendors that cleared the bar did so by being Ontario-based, not by being correct.