Already today, work of lawyers is being supported, if not fully automatized, by using artificial intelligence (AI). Background research is the first thing that is being done by computers, but eventually machines can make suggestions if not decisions on their own.
It is not automatization or software robotics that are the only drivers towards programmable legislation. We are already taking steps even further, including robots in our society with legal statuses. Robots too, will have responsibilities for their actions. And for that, they need to understand the legal ramifications for their actions.
More specifically, they need to be programmed to fully understand and abide by legal frameworks created by humans. This is especially true for robots that are self-governing, can learn and develop their operations autonomically.
Cost pressures will probably force us to automatize many aspects of government as well. This will also cause requirements for legislation, which is actually more like algorithms, guiding the work of robotized software governments.
Legislation therefore, needs to become compatible not only with “human interfaces” but also machine interfaces and AI. Hence the legislators of tomorrow may well need to understand programming.
We will truly be enhancing the concept of “programmable world”.
It will be interesting to see how AI and robotics impact things like intellectual property rights (IPR), insurances, criminal law and so on. What is the responsibility of the manufacturer or the owner of the robot? Is the software controlling the robot going to be made by a third party, is the robot “jailbroken”? What is the responsibility of whoever teaches anything to an AI entity and how can we control the contacts that an autonomous robot has?
As we saw with Microsoft’s AI chat bot Tay, things can truly go wrong with imperfect code. Especially when AI’s have other AI’s against them, bugs and vulnerabilities will quickly come up. Fujitsu has developed an interesting approach to customer service enhancing AI chatbots as part of Finplex offering, where there is extra focus on the control over machine learning. I wrote more generally about Fintech in my last blog.
Once we see software driving our cars, controlling robots that take our dogs out for a walk or make legal, political, financial and medical decisions, it all will become very real to each of us.
We need to be ready for the change, but most importantly, the code underneath these services – and legislation – needs to be ready.