It seems that 234 years after Immanuel Kant urged humans to have the courage to use their own understanding, the need to reason for ourselves is being removed altogether. As developments like artificial intelligence and machine learning find their way into every aspect of our business and social lives, our autonomy is steadily being replaced by the autonomy of systems.

The logic also holds for autonomous networking – the shift to a self-configuring, self-provisioning, self-healing and self-optimizing network – which is the widely accepted end game of network transformation. Cognitive networking is replacing manual processes. The network knows its capabilities and reacts to customer demand by providing the best services possible, while in the background resources utilization is continuously optimized. Bottlenecks are proactively identified, resources are added and activated in a fully automated way.

Despite the attractiveness of such long-term objectives, there are also some less positive side effects. Such fully automated networks will finally create a highly commoditized service. There will be no service differentiation any more other than pricing and the technical and operational headcount of service providers will continue to be reduced.

This creates two key challenges for today’s operators:

  • The business benefits of automation are too significant to ignore and so service providers have to transform their currently manually operated networks into fully autonomous network production factories. But this is complex; it requires a range of new skills and technologies and a new level of sophistication – as well as what will eventually be a much smaller work force.
  • As jobs disappear, new fields of work need to be identified. With value moving from the network to applications, it is quite likely that this will prove a very promising area – even if it does have very different skill requirements.

A vision of process automation and autonomous operations needs to be more than just a replacement of human interaction. In particular, it needs to find a way to value those whose ideas and creative drive deserve new challenges. Failure to address this might lead to an ever-widening gap between those who are actively engaged in value creation and those seeking entertainment.

You might argue that machines replacing manual work has been seen before, and that it turned out to create a lot of good. The difference these days is in the speed of change. Innovation is happening faster than people are able to assess impact. What’s more, there is no control through public discussion or governmental regulation.

Change is happening at unprecedented speed and it is quite likely that the established mechanisms for identifying and mitigating undesirable side effects will not be effective this time. Opinion-creating public discussion by civil society actors, academic leadership, governmental rulings and the engagement of trade unions all take too long and frequently only start when the consequences have become obvious. Those mechanisms will fail in our ever-faster moving, technology-driven world of continuous innovation.

This is why I believe, that the ones to develop innovation also need to actively engage by evaluating the social and economic impact in a more holistic way. They need to educate the public about potential consequences and predict impact. I do not see anyone else who can own this responsibility in times of rapid change with unprecedented impact on our societies and businesses.

A thorough technology assessment of the impact of autonomous systems is well overdue. How long can we as innovators ignore our responsibility to act?

If you’re interested in discussing this question further, please meet us in Paris at the AI Net Conference in the week of April 9. My presentation will explore the value of artificial intelligence in operationalizing SDN, and I’m certainly also interested in the wider discussion on the fascinating and urgent topic of AI.