The Industrialization of Artificial Intelligence presents new challenges to scientists, politicians and society in general
As we pointed out in the book Python Deep Learning, Artificial Intelligence (AI) can be considered as the “fourth industrial revolution”, as a way of highlighting the power of AI technology and its great impact in our society. This is because research in AI is going from an “academic-crafts” to “professionalization”, that we can refer to it as the Industrialization of AI. One representative example could be the maturity of the deep learning frameworks that has evolved from frameworks built at universities (Caffe, Theano) to industry-developed systems by Google (as TensorFlow) or Facebook (as PyTorch).
This causes that we are witnessing that corporate labs and private big datasets become more important in AI research. And that have devastating implications in public research because researchers in academia increasingly need to collaborate (and they become dependent) with the private sector to access the data and compute required to train state of the art AI systems. And this could (if it’s not already been done) skew this research or reduce its public value.
At the same time we are immersed in a wave of “democratization of AI” because the diffusion of AI research through open channels (as Arxiv), open-source software (as GitHub), open data and cloud computing available today to everybody, what makes it easier to deploy state of the art AI systems to any person. This implies important challenges for regulatory bodies who need to ensure the compliance in any AI environment, where adopting dangerous AI technologies is as simple as downloading some software from GitHub and installing it in any Cloud provider.
The industrial revolution resulted in massive social and political change, with consolidation of political power among capital-owners and the subsequent popular anger at the inequality this generated. It seems reasonable to think that the rise of AI follows the path described here, it will lead us to the same thing. We have to face what’s going on before it is too late. Otherwise we will face a disaster.