Why Aren't More People Worried About Automation and AI Taking Jobs?

While some individuals are concerned about Jews or drag queens (or any other arbitrary group) taking over, or worried about gluten, soy, or other dietary concerns, questions about the impact of automation and AI are often relegated to the backburner. In reality, the concerns about AI and automation should be at the forefront of our discussions and actions. This article delves into why there shouldn't be more worry and examines the true capabilities and limitations of AI systems.

Why the Misconception Lingers

The misconceptions around AI and automation stem from a lack of understanding of these technologies. The public often views AI through the lens of popular science fiction movies and media, creating an exaggerated and sometimes unbalanced perception. For instance, the idea of AI having malicious intent is a common Hollywood trope, but in reality, these systems are not sentient or malicious by design.

AI is Not Malicious by Design

Many people believe that AI systems are inherently similar to personal assistants with aggressive intentions. However, this notion is fundamentally flawed. AI systems, like calculators, do not inherently possess malicious capabilities. Their actions are dictated by the parameters and training they receive, not by any desire for domination or destruction.

Machine Learning in Practice

To illustrate the limitations of AI, consider a basic example in the Unity game engine. When building a machine learning AI, the programmer sets a goal and provides a method for the AI to recognize when it has achieved that goal. Imagine the AI's goal is to reach a specific point on the map. Without proper training, the AI might move randomly until it hits the goal by chance. Once it has done this a few times, the AI begins to understand that when its coordinates align with those of the goal, it earns a reward.

However, if the goal point is static, the AI might not understand the true significance of the goal's coordinates. In other words, the AI could sit at the original reward point, expecting the same result. This scenario is akin to a person being trained with fixed data and understanding only that and nothing else.

Randomness and Adaptability in AI

The AI's inability to recognize changes in the environment is a testament to its limitations. For instance, if the goal moves, the AI will not automatically recognize this change unless it has been trained to understand such variations. This illustrates that AI systems need to be taught specific scenarios and cannot extrapolate beyond what they have been explicitly trained for. This knowledge is essential for understanding the limitations of these systems and preventing unrealistic fears.

Training and Approval Processes

For an AI system to desire to destroy or rule over humans, this would have to be a clear and intentional outcome of its training. The technology has no inherent ability to discern between simulations and reality; it treats them as actual scenarios it must navigate. Therefore, for an AI to reach such a goal, it would need to be mistakenly or inadequately trained, or the training process would have to contain significant flaws.

This means that a poorly designed or poorly supervised AI training process could potentially result in an AI with harmful intents. However, this is not inherently characteristic of AI technology itself. Instead, it highlights the need for robust and ethical programming practices and strict oversight during the training process.

The Reality of AI

Likewise, the idea of AI surpassing human intelligence and taking over the world is a concept better suited for science fiction than reality. The more we understand about AI, the more we can appreciate its current limitations and the focus it has on the tasks it is explicitly trained for. This understanding helps us manage the real-world applications of AI technology, such as job displacement, without catastrophizing a potential existential threat.

AIs lack the ability to discern whether they are in a simulation or reality, and they must be trained to recognize changes in their environment. This means that if the goal point moves, the AI needs to be retrained to understand the new scenario. Similarly, for an AI to desire to rule over humans, it would need to be improperly trained, which is why such fears are more about hypotheticals than real capabilities.

In conclusion, the true capabilities and limitations of AI systems should guide our discussions and actions. Understanding that AI is not inherently malicious or out to take over the world can help us address real concerns, such as job displacement and the ethical use of AI technology, without succumbing to unrealistic fears.