Machines Like Us: Overcoming the Challenges of Coding AI with Common Sense

**Title: Machines Like Us: Towards AI With Common Sense**

**Understanding the Challenge of Artificial General Intelligence (AGI)**

Artificial intelligence (AI) has become an increasingly influential force in various industries. However, developing artificial general intelligence (AGI) remains a complex and difficult task. In “Machines Like Us: Toward AI With Common Sense” by Ronald J Brachman and Hector J. Levesque, published by The MIT Press, the authors provide a comprehensive overview of AGI and delve into the challenges associated with its development.

**Broadening Understanding of AGI**

The book offers a valuable resource for individuals seeking a broad understanding of AGI’s complexity. While it contains code and pseudo-code segments, they are straightforward and serve to support the text’s explanations. “Machines Like Us” not only benefits IT management by providing insights to critically evaluate vendors’ claims about AI solutions, but it is also suitable for anyone interested in the pervasive influence of AI across various aspects of our lives.

**The Importance of Common Sense in AI**

While computer software is proficient at executing known tasks and identifying exceptions, the authors argue that common sense is crucial for handling new and unfamiliar situations without the need for rewiring our brains. Common sense can be seen as a form of exception handling, which is imperative for AI systems to function effectively, even within restricted environments. The absence of common sense is cited as a prominent reason why fully automated vehicles are not yet a reality.

**Examining the Five Main Areas of AI**

The book concisely outlines the five primary domains of AI, serving as an excellent introduction to the subject. Though there may be room for interpretation, it offers a comprehensive overview of the current state of AI.

**Representing Knowledge and Common Sense**

The authors explore the topic of representing knowledge and common sense, emphasizing the significance of rules-based systems and expert systems, despite the latter falling out of favor. Their discussion acknowledges both the strengths and limitations of these approaches in replicating common sense. While deep learning is briefly mentioned, the book avoids excessive technical details often found in other texts.

**Integration of Deep Learning, Expert Systems, and Procedural Code**

Contrary to the belief that AI systems rely solely on deep learning, the authors highlight the effectiveness of combining deep learning, expert systems, and procedural code to comprehend data and provide intelligible information. Successful systems that utilize common sense for exception handling require a cohesive blend of these approaches and potentially undiscovered methods.

**Implementing a Common Sense System**

Chapter nine emerges as a pivotal chapter, offering insights into implementing a common sense system. After exploring common sense, AI systems, and representational constructs, the book emphasizes the need for a blended approach that combines top-down and bottom-up methodologies. Rather than advocating for a single solution or downplaying the difficulties associated with AGI, the authors emphasize the importance of exploring multiple avenues to expand our knowledge.

**Embracing the Complexity of AGI**

The book reminds readers that AGI is a multifaceted challenge that necessitates further exploration. It challenges the binary perspectives often expressed: those who believe AGI is insurmountable and those who overly rely on their own tools as the ultimate solution. With the human brain still not fully understood and acknowledged as a significant source of intelligence, the authors endorse the pursuit of diverse techniques to advance our understanding of AGI.

**Conclusion**

“Machines Like Us: Toward AI With Common Sense” is a valuable resource for individuals seeking a comprehensive overview of AGI and its challenges. By shedding light on topics such as common sense, knowledge representation, and the integration of various AI approaches, this book contributes to a deeper understanding of the complexities involved in developing truly intelligent AI systems.