Google Health’s claim in Nature that its AI program had outperformed professionals in diagnosing breast cancer is suspect: “In October, a group led by Benjamin Haibe-Kains, a computational genomics researcher, criticized the Google health paper, arguing that the ‘lack of details of the methods and algorithm code undermines its scientific value.’” The problem is, the details are in the code and Google won’t share the code. For example, many readers may be surprised to learn this item from his recent Scientific American article: Today, while AI appears to be booming, Horgan says, hype frustrates critical appraisal of advances. By 1998, problems like non-recurrent engineering had begun to be recognized: “Algorithms that can perform a specialized task, like playing chess, cannot be easily adapted for other purposes.” John Horgan, “ Will Artificial Intelligence Ever Live Up to Its Hype?” at Scientific American (December 4, 2020)īut that year, 1984, ushered in an AI winter, in which innovation stalled and funding dried up. I edited an article in which computer scientist Frederick Hayes-Roth predicted that AI would soon replace experts in law, medicine, finance and other professions. IEEE Spectrum, the technology magazine for which I worked, produced a special issue on how AI would transform the world. When I started writing about science decades ago, artificial intelligence seemed ascendant.
At first, science writer John Horgan (pictured), author of a number of books including The End of Science (1996), accepted the conventional AI story: