David Walton's newest novel, Three Laws Lethal - title inspired by Isaac Asimov's three laws of robotics - begins with what certainly is an ethical quandary that typifies our increasingly AI-driven age, in this case, driven literally. A mother with her children are passengers in an AI-driven automobile. She can turn around and tell them to stop arguing, without risking an accident. She marvels at being in the driver's seat with her hands off the wheel. And then ... A big tree falls in front of them. To plow into the tree would risk the death of both mother and children. The AI computes the deadly odds, and acts upon it, instantly swerving the car to the right to avoid the tree. Unfortunately, there's a biker in that lane, and he's killed by the swerving car.
It's not that the AI didn't see the biker - the problem is that it did, and decided the mother and children's lives were more worth saving than the biker's. Now, people driving in our reality make split-second decisions like this all the time. They're maybe not quite decisions but instant gut reactions. Would anyone for a moment think of charging a mother with vehicular homicide if she did what the AI did in the car? Of course not. But there's something deeply disturbing about an AI making this decision, any decision, that results in the loss of innocent human life.
This is the problem that opens David Walton's novel, just published by Pyr today. It's a narrative that is as philosophically profound as it is breathtaking. Asimov imagined/foresaw that all robots could and would be programmed with three laws: 1. A robot can never do harm, or allow harm to be done, to a human being. 2. A robot must follow all orders given to it by a human, except when following such an order would contradict the first law, i.e., harm or allow harm to be befall a human. 3. A robot must always act to protect itself, except when that would contradict the first or second law. Asimov wrote great novels and stories that explored what could happen when these laws were bent or broken. In that sense Three Laws Lethal is an extrapolation of Asimov, a meditation on how an AI programmed to protect human lives can end up taking a life - a life that threatens no one, but whose existence nonetheless must be ended to protect the people the AI serves. The novel is also Asimovian in the sense that it is an un-put-downable read.
Exploration of driverless cars makes Three Laws Lethal not a story happening the day after tomorrow, as they used to say back in the 1950s, but maybe more like a few minutes from now. An Uber driverless car already killed a woman in Tempe, Arizona in 2018. This apparently was more a malfunction than a deliberate decision of the car to kill the woman (a pedestrian walking her bicycle in the street), so it didn't raise the kind of wrenching ethical dilemmas posed by Walton. But Walton also explores, with the same panache and savvy, the corporate competition and intrigue that has characterized the digital revolution since it began in the 1980s. In this case, we go from intrigue to outright assassinations, and self-driving cars to fleets that can work as an attack force. Malignant AIs reminiscent of HAL (an Arthur C. Clarke not Asimov creation) also figure in this story, pitching it at least a few years and maybe decades into the future.
All of this is played out by a memorable cast of characters all along the continuum of fundamental human decency, which at the bad end includes a willingness to do the aforementioned murders to get desired results. In as much Asimov's robot stories were also detective stories, this makes Three Laws Lethal an Asimovian story in yet a third, appealing sense.
Although Asimov defined the genre of sentient robots and therefore AIs, the other two titans of the golden age of science made important contributions to this crucial sub-genre. In addition to Clarke's homicidal HAL, Robert Heinlein's self-sacrificing Mike has a permanent place in the AI pantheon. No one can duplicate those achievements, but it's good to see that David Walton is carrying forward that tradition so well as we move to ever more AI in our cars and lives in the 21st century.
No comments:
Post a Comment