The Human Perspective in AI Development

Zena Assaad, a senior lecturer at the Australian National University’s College of Engineering, Computing and Cybernetics, focuses her research on human-machine teaming, robotics, autonomous systems, and AI.

Science fiction often portrays artificial intelligence as a reflection of human intelligence, yet the reality diverges significantly. Initially, AI was conceived as the replication of human intelligence in machines, sparking debates on the semantics of intelligence.

Human intelligence is a subjective and complex concept, making its replication in machines challenging. Software, being binary, fails to capture the nuances of human thought and decision-making.

AI systems may have goals, but these differ from human intent. Goals are specific and measurable, while intent involves underlying purpose and motivation. AI lacks the capacity for intent and reasoning, complicating the replication of human intelligence.

Discussions on AI ethics often overlook the fact that technology evolves under human direction. Incorrect outputs from AI systems result from design flaws, not unethical behavior by machines.

Achieving ethical AI requires embedding ethical decision-making throughout the AI supply chain, from data labelers to consumers. Everything has a cost, and ethics concerns the costs we are willing to bear.