You are here:
- Home
- AI – Applied Intelligence
AI – Applied Intelligence
A unique Artificial Intelligence capability – the first able to understand people, the real world, and cause and effect – that does not rely on data, statistics, or natural language processing.
Unlike traditional statistical systems, this AI adapts to individual situations in context. By simulating these, it is able to generate and optimize courses of action, explain why, predict the most reasonable future, and create adaptive plans that accurately take current conditions into account.
“…any AGI is useless unless it can be trusted, which means that systems must be able to understand the world well enough to justify their conclusions, provide causal explanations, prove that they are correct, and help others understand why various paths are correct.”
“...everything in the real world takes place within, and cannot be separated from, some context. The same action taken in different contexts will generate radically different effects, which makes success and failure entirely dependent on context. This places context at the core of goal fulfilment, which in turn makes it critical to intelligence. Differences in context also lead to differences in explanations; people can only act with confidence when they understand why things are as they are, and in significant part it is a system’s deep understanding of context that enables it to provide this understanding.”
“Only systems that can be actually proved to be safe and correct can be trusted enough to actually be deployed in practice. If a system could possibly fail in unpredictable ways, it’s the same as if it never existed at all – you wouldn’t be able to rely on it enough to give it the autonomy it needs to be useful. Autonomy is required, however, in order to compete with others’ superintelligent systems. Moreover, in deploying unsafe systems we risk consuming all available resources trying (and ultimately failing) to control them.”
“...everything in the real world takes place within, and cannot be separated from, some context. The same action taken in different contexts will generate radically different effects, which makes success and failure entirely dependent on context. This places context at the core of goal fulfilment, which in turn makes it critical to intelligence. Differences in context also lead to differences in explanations; people can only act with confidence when they understand why things are as they are, and in significant part it is a system’s deep understanding of context that enables it to provide this understanding.”
“Only systems that can be actually proved to be safe and correct can be trusted enough to actually be deployed in practice. If a system could possibly fail in unpredictable ways, it’s the same as if it never existed at all – you wouldn’t be able to rely on it enough to give it the autonomy it needs to be useful. Autonomy is required, however, in order to compete with others’ superintelligent systems. Moreover, in deploying unsafe systems we risk consuming all available resources trying (and ultimately failing) to control them.”
Daniel Olsher
Integral Mind 2024