GDJ / Pixabay

The recent ISC 2017 event saw the unveiling of a high-performance architecture that combines the powers of High Performance Analytics (HPA) and High Performance Computing (HPC). This architecture is driven by an Exabyte-scale memory capacity.

To put it in simple terms, the world became witness to what will drive the future of AI by bringing into existence the first scalable learning machine. It will be trained using data from ages gone by and bring us closer to the reality of Advanced AI. In other words, we are looking at something that will outdo humans in every way possible.

High-Level Machine Intelligence (HLMI) will drastically change the entire way the current world functions. In a century or so, we could see all human work being handled by machines.

A recent white paper published by researchers at Yale University, Oxford’s Future of Humanity Institute, and AI Impacts, predicts that HLMI will even be capable of creative work.

The paper outlines two particular watershed moments that will have a revolutionary impact as far as AI is concerned. The first will be the creation of HLMI and the second will be a phase called “Intelligence Explosion”, which will be facilitated by HLMI.

What to expect

AI, at this level, will transform everything from health care to military. In order to adapt, global leaders will have to shape public policy that is capable of adapting to this kind of change.

Millions of jobs will be taken over by machines and not only will this lead to a spike in unemployment but also bring in challenges related to infrastructural requirements, laws and regulations, and cyber-security.

The researchers also carried out a survey that questioned around 21 respondents; all of them researchers who had their work published at the International Conference on Machine Learning (ICML) and 2015 Neural Information Processing (NIPS) conferences.

Some of the key questions that came up as a result include:

  • Whether AI development would progress at a rapid rate after the automation of AI development itself?
  • How would HLMI impact the economy?
  • What should be done in order to derive positive outcomes from AI Progress?
  • What are the chances of extreme outcomes- both, positive and negative?

The authors of the survey believe that cost-effectiveness of tasks/processes will be a key factor in determining the outcome of an AI “takeover”. If the tasks are at lower costs and with greater efficiency, we could be looking at tremendous consequences from a social standpoint.

However, many of the respondents, who are AI Experts, believe that the impact would lean towards being more positive than negative. 45% of respondents agreed to the possibility of positive outcomes, while only 10% and 5% agreed to negative and extremely negative outcomes, respectively. The Extremely negative outcome here refers to human extinction.

However, the authors of the study saw it necessary to point out that, irrespective of the kind of outcome, societies and governments must find a way to inhibit the risks posed by AI progress.