                                                                     A CRITIQUE OF THE STANFORD 100 YEAR STUDY ON AI

    The Stanford One Hundred Year Study on Artificial Intelligence is an ambitious project to chronicle the evolution of AI technologies. The first report covers a plethora of issues concerning AI technologies affecting many fields of human endeavor. 

       In this critique, I challenge the predictions and recommendations of the Report related to the autonomous or 'self-driving' technologies on two fronts. Firstly, there is no mention of the need to have one common design philosophy concerning the self-driving tech. Secondly, the assumption that gaining public trust is sufficient to ensure widespread acceptance of this technology. Let us examine this one by one.

       At present, major automakers and multinational corporations are pouring billions of dollars into research and development to be the first to field self-driving cars and get the early bird advantage' in the field. While in a way this competition is good for the fast progress of the technology, the lack of a coordinated effort in developing self-driving tech may have detrimental effects. The design philosophies may vary, and therefore the characteristics and behaviors of cars from different manufacturers will differ wildly. When we have a number of these cars together jostling for space and customers on the roads, the possibilities of mishaps are many. For example, in initial phases, some manufacturers will involve a human driver as a backup for unforeseen circumstances, whereas others will adopt a different strategy. This mix of design philosophies with no guidance of which one is the right' approach, can be quite dangerous and hamper progress towards adoption. A concerted effort towards a common design philosophy of self-driving tech is a more logical approach that needs to be adopted by all stakeholders. 

       Secondly, the Report assumes that the adoption of self-driving tech will be governed mostly by public trust in the technology and that they may eliminate the need to own cars. The report predicts that the technology will be highly pervasive and its impact on the transportation sector will be total. However, no emphasis is laid on the crucial Transition period' between the time when the technology is first introduced and its ultimate universal adoption. During this period, we may have human drivers along with self-driving cars on our roads. Many factors affect the duration of this transition. One of the factors is the affordability of the self-driving tech. If the common man cannot afford a self-driving car, then its adoption among the masses will be slow. Also, a lot depends on the timeframe for a complete overhaul of the public transportation system with this technology. Even if we account for all these factors, there may still be hobbyists and enthusiasts' who may be averse to the idea of shunning traditional vehicles just because of the sheer joy of driving. At some point in time we may have to consider the question: Do we ban human drivers from our roads and stop production of such vehicles? 

       Policy makers need to consider above factors and formulate of laws governing such a mix of vehicles on the roads. The SELF Drive act passed by Congress recently is a good step in this direction. This bill gives powers to the National Highway Traffic Safety Administration (NHTSA) to regulate the design, construction and performance of self-driving cars within a common framework. As per the act, to get a permit, a self-driving car needs to pass tests which are formulated by NHTSA to be deemed as safe as a human driven car and therefore fit for the roads.
