Charles Darwin in 1859 set the stage for the emergence of artificial intelligence (AI) when stating in the final sentence of his book Origin of Species, “from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved.” According to an article with the title, “From So Simple a Beginning:” Species of Artificial Intelligence that was published in the Spring 2022 issue of the journal Daedalus, artificial intelligence has a decades-long history exhibiting alternating enthusiasm and disillusionment for the field’s scientific insights, technical accomplishments, and socioeconomic impact. Recent achievements involve renewed claims for the transformative and disruptive effects of AI. Exponential increases in computing power, open-source software, available data, and embedded services have been crucial to this success. Yet, there is growing unease around whether the behavior of these systems can be rendered transparent, explainable, unbiased, and accountable. The author argues that artificial general intelligence (AGI)—able to range across widely differing tasks and contexts—is unlikely to be developed, or emerge, any time soon.
From driving cars to controlling critical infrastructure, from diagnosing illnesses to recommending content for entertainment, AI is ubiquitous. When in 2011 IBM announced a new age of cognitive computing with Watson, it was asked, why not turn Watson into a physician, but task transfer and generalization have turned out to be quite difficult. A physician’s general problem-solving is full of task and context changes. Rather than replicating accomplished physicians, IBM’s Watson Health has turned out AI assistants that can perform routine tasks. Recent possession of symbolic language and discovery of mathematics and formal systems of computation have provided tools to build and explore new AI systems, a broad repertoire of approaches and methods that remains essential. AI systems with their ability to represent and discover patterns in high dimensional data have as yet low dimensional embedding in the physical and digital worlds they inhabit. This thin tissue of grounding, of being in the world, represents the single largest challenge to realizing AGI, systems able to range across widely differing tasks and contexts reflectively.