Stanford researchers participating in the study AI100 (aka One Hundred Year Study on Artificial Intelligence) has just published their report on the impact of artificial intelligence and robotics on our society by 2030. The report is reassuring, it highlights a number of dangers and made a series of proposals to deal with the imbalances that are sure to bring progress in artificial intelligence.
AI100 Committee researchers will want reassuring
A scenario of type Terminator is purely imaginary, plausible or even inevitable while accelerating progress in artificial intelligence and robotics? This is the theme of the study AI100. If the study group is officially based at Stanford, members of the Committee to study go far beyond the University of California then that there are several other major universities American, as well as Eric Horvitz, the Director of Microsoft Research or Astro Teller, the Director of Google X.
The results of their study are rather reassuring. These researchers believe that advances in artificial intelligence are enough progressive for the man has time to adjust. Researchers estimate that applications such as autonomous driving, with the reduction in the number of accidents on the road expected, should help the population to better accept the artificial intelligence in their daily lives. Human intelligence and artificial intelligence should be able to complement one another. Nevertheless, researchers point to a number of risks. On the one hand the dissemination of AI in society can inequality between those who have access and others. Similarly, the risks relating to the protection of privacy are evident in an environment where the machines collect data continuously to feed these artificial intelligences. The smartphone is the most glaring illustration of this state of affairs that will grow again with the introduction of ubiquitous voice assistants and the imminent arrival of robot companions.
3 recommendations to prepare for the spread of AI
At the conclusion of their study, the researchers deliver 3 recommendations. On the one hand, they believe that Governments and legislators must get in expertise on the subject of intelligence artificial and well understand the technologies and their impact before legislating, block or allow any innovation in this area without a good understanding of the issues. The researchers estimated that an effort must be made as to the analysis of the IA, including those who have a direct action on the physical world. We need to conduct a reverse engineering of proprietary systems to understand their operating rules so that they can be discussed by the research, academics and journalists so that they will be accepted or, if applicable, denied by the company. We think especially of the now famous macabre dilemna of the autonomous car. Finally, researchers call for more resources (public but also private) so that interdisciplinary studies can be initiated in order to better evaluate the involvement of societal of AI on the world of 2030.
Translation : Bing Translator
“One Hundred Year Study on Artificial Intelligence (AI100)“, 2016 Report
“Stanford study concludes next generation of robots won’t try to kill us”, DigitalTrends, November 27, 2016