Can AI determine where i am in 10 yes or no questions ?

The user is asking whether a classic guessing game—narrowing down a location with ten yes-or-no questions—can actually succeed. While the game is intuitive, the mechanics of real-world geolocation raise questions about feasibility. We’ll examine what such a strategy would realistically entail.

Background

AI systems can generate yes/no strategies for approximating a person’s location, but achieving global, real-time accuracy with only ten questions remains beyond the reach of today’s open-source tools. Practically, any such game operates on a pre-filtered set of candidate locations rather than discovering coordinates from scratch. Systems typically begin with coarse geolocation databases (e.g., IP-to-region mappings, language inference, known Wi-Fi SSID clusters) to shrink the search space to hundreds or thousands of possibilities before the conversation even starts. From that point, yes/no questions may further winnow the list, but the initial coarse filter is indispensable; without it, the number of viable candidates would remain far too large to exhaust in ten questions. Commercial assistants (e.g., voice-activated smart speakers) prioritize fast, low-cognitive-load interactions over minimizing question counts, so their implementations may spare users the full ten-question sequence even when a precise answer is required. These trade-offs between speed, robustness, and question economy highlight why the “ten-question” figure is best understood as a heuristic for ranking small candidate sets rather than a universal discovery protocol.