In May of 2016, the then-startup Otto had a self-driving semi-truck travel 120 miles on a Colorado freeway to deliver 2 000 cases of beer. Uber bought Otto in August of that year, and, together with Volvo, has been working on technology for self-driving vehicles since then. In late 2017, Uber quietly started running its small fleet of self-driving semis on Arizona highways and, in early March of this year, it officially inaugurated Uber Freight, its service using these trucks.
Then, on March 18, an autonomous Uber test car killed a pedestrian at a crosswalk in Arizona. Uber immediately cancelled all of its testing in various cities, and on March 26, Arizona banned Uber vehicles from its roads. According to Arizona’s governor, Uber had not put citizen safety as a priority and would no longer have the right to test in the state.
Tragically enough, there had been a “safety driver” in the car at the time who was supposed to oversee what was happening and take over in case of an emergency. Uber’s self-driving trucks also have a licensed truck driver on board for the same reasons. In this instance, the artificial intelligence that powered the vehicles was not advanced enough to operate with complete autonomy, and so a human presence is still required. We are still a ways off then from entirely self-sufficient vehicles operated by artificial intelligence, but not that far. Human fatalities cannot be a part of the testing process, but while we wait for the technology to be improved and regulations to be considered, we do need to ask ourselves what we are getting into. What exactly is artificial intelligence, and what will its impact be?
Artificial intelligence (or AI) is a term that describes the concept of machines, in particular computer systems, displaying intellectual characteristics typical of humans. Reasoning, voice recognition, decision making, and learning from past experience are all examples of abilities which have been developed in AI.
AI has reached a certain level of sophistication that is making a broader integration of its technology into our daily lives more of a reality and less science fiction. However, what we don’t always realize is that we are already living with certain applications of the technology and some of it for quite some time. Let’s take a look at a few examples here:
Autopilot: In 1903, the Wright brothers had their first sustained flight of just under a minute. Commercial flying took some time to really take off (pun intended) but auto pilot did not. In 1914, Lawrence Sperry demonstrated the first device that would allow a plane to fly on its own, and since then the principles of the system have been the basis for most autopilots in use. It is estimated that in most longer-haul commercial flights, a pilot is “hands on” for approximately seven minutes only, which is roughly the time needed for take off and landing. The rest of the flight is conducted through autopilot.
Google Maps: the suggested routes of Google Maps are based on an instantaneous analysis of location data from smartphones as well as user-reported traffic incidents such as construction and accidents.
Siri: when you ask Siri a question and she answers back, a highly sophisticated system of voice recognition and speech analysis is it work. Your speech enters the computer and is compared against a statistical model to estimate which letters you may have used in order to determine roughly which ones were vowels or consonants. Based on this, your speech is run through a language model that will estimate the words spoken, and the computer (through Siri) will then give you a list of possible interpretations. Once you say which interpretation is correct, Siri will respond with an answer.
In these examples, a computer has been programmed to analyze all sorts of different data sets and then choose an action (auto pilot), make a decision (best possible route on Google Maps), or understand and respond to a spoken query, giving a list of choices (Siri, who also happens to be very polite). All of these responses to a problem require reasoning, consideration, and analysis, and it is all done instantaneously.
Very soon, the once-futuristic seeming idea of a self driving car will become reality. There is already AI that is able to perform certain kinds of legal analysis at an intermediate level. It has entered into the financial realm, plowing through data and offering up different strategies based on stated goals. AI has clearly reached a point of advanced comprehensive ability with a knowledge base deep enough to use as reference. The depth of its capabilities is astonishing, and its presence in different aspects of our daily lives will only continue to grow.