The question of how AI will integrate with our society and economy is a fascinating one. We often make the mistake of assuming that an AI will be similar to a h

The question of how AI will integrate with our society and economy is a fascinating one. We often make the mistake of assuming that an AI will be similar to a human just faster or smarter, but that misses some of the key distinctions of an AI versus biological intelligence. One of the most striking is the ability to radically alter the substrate and operation of an AI system. Because of the emergent nature of intelligence, I suspect that many AI instances will be raised like children, tested and validated for specific environments and then large portions of their consciousness could be frozen to prevent divergence of their operational modes. AI systems could also incorporate self-auditors, semi-independent AIs which have been raised to monitor the activities of the primary control AIs. Just as we involve checks and balances in corporate or national governance, many AIs may be composite entities with a variety of instances optimized for different roles. This will be desirable since you may not want a general AI intelligence acting as a butler or chauffeur. Do you really want them to be able to develop and evolve independently? Of course this just scratches the surface. AI will take in us in directions we can not dream of today.