There are several different paths for AI progress. There’s the Theory of Mind. It’s a theoretical system that grows and evolves, learning from past experiences and simulations to predict outcomes. It’s important to realize that AI can’t replace human intelligence completely. While it can perform many tasks, it’s still not as powerful as the best human. Here’s a look at some of the ways that AI might eventually replace humans.
Rather than being an agent seeking a goal, AI could be an algorithm. Nanotechnology scientist Eric Drexler, author of Engines of Creation, argues that AI should not be viewed as a single agent. Instead, AI should be categorized into Comprehensive AI Services, a series of subroutines that will do specific tasks. In this way, AI could do everything from solve complex problems to make predictions.
Another alternative to artificial intelligence is Organic Intelligence. This method of learning is based on the principle that human beings learn from experience. By learning from data, an organic system could learn to perform cognitive tasks as well as humans. The goal of this process is to develop a machine that is similar to human intelligence. But what exactly is the difference? And what’s the difference between artificial intelligence and organic intelligence?
Before the invention of AI, humans had to struggle to create intelligent robots. Aristotle had a theory that described how the brain learns to learn and act. Hebb’s concept of “hebbian learning” continues to be a valuable model for AI. And then there’s the Turing Test. These two theories are at the forefront of AI research. This is a paradigm shift in the tech industry.
When it comes to detecting fraud and disease, timing is everything. Identifying an outbreak’s symptoms requires early warnings from networks of citizens, medical practitioners, and experts. Using AI, distributed human intelligence and expert human intervention can help coordinate response. A good example of this is the emergence of COVID-19. In the meantime, there’s no magic COVID-19 solution for detecting human fraud. Instead, the solution to COVID-19 lies in a multifaceted, interconnected network of human experts who can interpret big data.
The Ripple-Down Rules approach to AI addresses these problems. It suggests that the problems of conventional methods stem from philosophical assumptions. In other words, the problem of knowledge acquisition comes from a faulty frame of reference. People never explain how they reached a conclusion. Instead, they justify it with examples and differentiating between cases. RDR focuses on a more contextualized understanding of knowledge. The paper provides numerous examples of RDR using free software.
The latest developments in deep learning are likely to continue to have a major impact on AI. However, despite its nascent success, deep learning still has not shown strong reasoning. This may change with the next generation of AI. A better way to train machines is to use small amounts of data. By combining small datasets with unsupervised learning, we can improve the quality of AI. If we can achieve that goal, we can help humanity and our planet.