Johannesburg, 14 Feb 2016
Whether we like it or not, artificial intelligence ("AI") is here, and it is here to stay.
AI development is progressing rapidly and it is predicted that AI will become mainstream in the next few years. There appears to be little doubt that AI will have an unprecedented impact on the way we work, live and play - if not our very existence.
This article briefly identifies some of the potential benefits and opportunities that AI presents and raises some of the key challenges and legal risks that are likely to follow.
AI in a nutshell
AI is a broad concept referring to the "intelligence" of machines, computers and software. It is more than just robots or what we've seen on the big screen. We use AI in our daily lives already (e.g. Siri or Google Maps), and some dramatic 'futuristic' AI applications are also still in development.
There are three major categories for AI:
(1) Artificial Narrow Intelligence (ANI) or 'Weak AI': This is where a machine only specialises in one area. This is the kind of AI we commonly see today.
(2) Artificial General Intelligence (AGI) or Strong AI: This refers to machines that are generally as intelligent as humans across all aspects. In other words, these machines can perform any function that a human can. This has not yet been developed.
(3) Artificial Super intelligence (ASI): An ASI would be able to perform more intelligently than the best human brain in practically every field, including scientific creativity, general wisdom and social skills. While not yet developed, this type of AI is at the core of aspirational AI research.
So the aim is for AI to not only mimic human intelligence, but to ultimately tsurpass the intelligence of even the brightest of humans.
Reasons to get excited
We've already read about drones, self-driving cars and engaged with Siri and Google Now. IBM Watson (a computer system capable of answering questions posed in natural language and revealing insights from large amounts of unstructured data) has been integrated into e-commerce Web sites to predict our shopping patterns. We can imagine how robots should increase productivity and therefore the profitability of companies, with resultant positive effects on the economy. Proponents from the health and sciences fields foresee AI technology applications that include implants that enhance brain functioning and nanobots that increase oxygen flow to the human body.
But this is only the beginning. Advocates of AI argue that humans will amplify their own capacity by merging their skills and resources with those of their intelligent creations, and that the potentially powerful applications of AI are endless.
Stephen Hawking, Elon Musk, Google's director of research Peter Norvig and about 150 other prominent scientists and industry experts published an open letter in January 2015 on artificial intelligence. They contend that:
"[t]he potential benefits are huge, since everything that civilisation has to offer is a product of human intelligence; we cannot predict what we might achieve when this intelligence is magnified by the tools AI may provide, but the eradication of disease and poverty are not unfathomable..."
Should you be worried?
Hollywood has painted a grim picture of what a future with AI will be like. From the Terminator franchise to Ex Machina, iRobot and the latest instalment of the Avengers, it would appear that this could possibly be our best and last invention.
Although they actively support research into artificial intelligence, Hawking et al's open letter also makes it very clear that they fear the potential pitfalls of AI. Their fundamental long-term concern relates to control and imposing limitations on machine autonomy: machines should be designed to ensure that they remain susceptible to our control and amenable to our wishes, failing which it could threaten humanity.
And could you lose your job? Perhaps. Some argue that most of the jobs held by humans today can be performed more effectively, efficiently and cost effectively by robots - they don't get sick or tired or require salaries, benefits, training or career growth. Others have a more optimistic view that artificial intelligence will have a similar impact to industrialisation and the Internet on our economy: while some jobs will be replaced by robots, new jobs will also be created that will be filled by humans. It's probable that artificial intelligence will change the face of the economy and leave some (or many) people displaced, but the jury is out on how soon this will be.
Legal Challenges
South African legislation already circumscribes the requirements and conditions for the lawful processing of personal information. One implication is that machines operating in South Africa must be programmed and controlled to ensure that any processing of information is performed within the parameters of the law.
However, South African law does not yet address all the potential issues that might arise from artificial intelligence. We must therefore consider what the laws on AI might cover.
Lawmakers are often guided by public policy considerations. There are also some obvious ethical issues and questions, which are likely to arise in the context of AI. A few examples follow.
* Should the law regulate the level of control that should be exercised by humans over AI machines?
* At what stage, if any, should AI intelligent machines be considered legal persons? What would be the threshold for personhood - e.g. would it depend on the level of intelligence and autonomy?
* Who should be culpable for any misuse or malfunction of an intelligent machine, and in what proportion?
* Should machines be permitted to make decisions that affect whether people live or die or are injured (for instance, in the case of self-driving cars making decisions where there are strong possibilities of accidents)? And should anyone be liable in the event that someone dies or is injured?
These are only a few issues that warrant consideration from a legal perspective in the context of AI. We look forward to watching (and hopefully influencing) how the law develops in line with this interesting technology.
Did you know?
AI is made possible by the convergence of at least four fast-advancing technologies:
(1) Machine processing: The ability of machines to process text, images, speech and languages.
(2) Machine learning: The ability of machines to go beyond what they are programmed to do and actually teach themselves. This involves using algorithms to process large datasets in order to mirror the functioning of the human brain, particularly the ability to adapt to new circumstances and to detect and extrapolate patterns.
(3) Machine perception: The capability of a machine to use its senses (vision, hearing and touch) to interpret and relate to the world in a similar way to humans
(4) Machine control: This relates to the improved design of machines involving better control mechanisms with the ultimate goal of enhancing the speed, sensitivity and accuracy of AI responses.
Share