‘Mapping the contours between humans and machines is becoming one of the most intriguing, and at times creepy, challenges of our times,’ writes John Thornhill, the FT Innovation editor. ‘You shouldn’t anthropomorphise computers because they don’t like it,’ he quips. His interaction with a robot called Sophia which (or ‘who’) possesses ‘mesmerising lifelike facial features’ shouldn’t be viewed as anything more than ‘technological tools or digital slaves designed to do our express bidding.’ The ‘cutesy human stuff’ is nothing more than ‘false advertising’. Today, robots function as security guards, nursing assistants, teachers and sex toys; but a decade hence, they will be smarter, and could even include some moral agency. ‘Algorithms in self-driving cars, for example, already indirectly involve life and death decisions, prompting some car companies to employ philosophers to devise ethical settings in their systems.
Understanding how artificial neural networks make judgments is important particularly for doctors and the military but also companies. According to Richard Waters, the FT’s US West Coast editor ‘even the experts cannot tell exactly why robots come up with the answers they do.’ He points to the Tesla car driver killed when the ‘autopilot’ software failed to identify a white truck on a sunny day. ‘It’s a big black box — it can’t engage in a conversation with you.’ Researchers are now looking to employ teachers to train AI systems as if they were normal students, starting with simple concepts. It’s an approach that might work well for the human/AI workforces of the future.
Tim Hartford, the FT’s economics leader writer, is less fearful of AI’s impact on jobs. He believes that it will enhance rather than replace human activity, improving pay and creating more interesting work for humans. Indeed, his concern is that the AI revolution isn’t happening fast enough, and writes mockingly: ‘Sorry: this robot takeover could not be completed at present.’
In another, fascinating article, ‘What we get wrong about technology’ Hartford argues that ‘the most influential new technologies are often humble and cheap.’ It wasn’t the Gutenberg press that changed the world as much as the invention of paper on which it printed, giving rise to everything from wall decorations to toilet paper. Similarly, the invention in 1874 of barbed wire enabled American settlers to protect their crops from roaming bison, while the simple shipping container, invented in the 1950’s, transformed global trade.
In our rush to understand AI, corporate leaders should not overlook the modest innovations whose impact might be just as dramatic.