Artificial intelligence across everything we make
BY SANDY STRASSER
(Published in The Produktkulturmagazin issue 2 2017)
At the beginning of the year, “Europe's hottest conference invitation”, the DLD conference, brought together the most influential opinion-makers, industry leaders, start-ups and digital giants once again. Founded in 2005, the conference has developed into an interdisciplinary, internationally connected platform around the globe with top keynote speakers. Microsoft CEO Satya Nadella for example shared how Microsoft is democratising AI for everyone during his keynote in January in Munich. In an interview, he also talked about how we can use artificial intelligence to make the world a little better.
Mr. Nadella, when will your company be replacing you with a robot?
There was a McKinsey report on all kinds of jobs and their potential for automation. The conclusion was that around 30 percent of those jobs at management level could be automated. I thought that that is exactly what I need – because I sometimes have difficulty remembering what day it is. And if I don’t know what day it is, the 30 percent of additional time provided by a robot certainly couldn’t hurt. With artificial intelligence, we are actually trying to teach machines things that us humans do, but always in a way that ultimately benefits us. It’s basically about improving what we already have.
In general, we are still in a mainframe error. After a kind of AI winter, we are now experiencing a surprising AI renaissance. Due to the masses of data – and most notably label data and smart algorithms – we are currently in the midst of an AI and machine learning resurrection. New machine learning technology called ‘deep-neuro maths’ has been developed, particularly over the past five years. It has decisively influenced our understanding of our perception. This is truly phenomenal. However, if you ask me whether we have somehow come closer to a kind of general artificial intelligence, then I would have to say no. In the next phase, I believe we must concentrate on democratising access to AI. Instead of admiring those six or seven large enterprises who are majorly focusing on AI, we must start understanding that AI is omnipresent and in all companies – in all interfaces and in every interaction between man and computer. Everything is AI-controlled.
How do you define AI in comparison to general artificial intelligence?
When we simulate our brains and the decisions that our brains make and we transfer the acquired learning from one area to another, then this approximates AGI – or artificial general intelligence.
You state that there are only a few businesses that are majorly focusing on artificial intelligence. Please describe your view of the current competitive landscape and how the players differ in the battle of the brains.
Initially, we should try to better understand the identity of these companies. What do I expect from Microsoft and what do I want from Facebook? What can Google or Amazon do for me? When I look back at our own successes and failures, I think that we were right to not constantly be peering over the fence to see what others were doing – even if we would have been able to learn something from their skills and capabilities. It is of course not wrong to admire the skills and capabilities of others, but you have to apply them only to your own identity and to the expectations that customers have of your own company.
What is the most important factor in such competitive situations?
The history of our industry has always been determined by whatever there has been a shortage of. And this frequently then becomes the merchandise of tomorrow. Today, we all talk about competition regarding talent and data. I see the race like this: how can I create new value from the current situation from existing products? This is how our sector functions. Sometimes, you do it right and you are successful – sometimes it fails and you then have to wait until the next round. This is what competition is.
How do you decide when and in which form you replace human intelligence with computer technology?
On the one hand, I think that we require new technologies and innovations that, above all, drive growth. On the other hand, technological progress results in unemployment, and getting that under control is probably one of the greatest challenges. Therefore, we have to return to what economists term the ‘lump of labour fallacy’ when talking about the Industrial Revolution. This means that not just the technology sector, but also all other areas of industry, the public sector and governments, must become involved in order to prepare and train people for the jobs of the future. I am therefore thrilled by what Germany has achieved since its reunification and by what Switzerland has done with regards to its training programmes and internships and how jobs can be created through these. This has to be replicated across the globe. A new balance has to be achieved whenever you have considerably higher revenues generated from capital than from manpower. This was also the case during the Industrial Revolution, which brought about the labour movement, which in turn created the social network that continues to benefit us all to this very day. Therefore, we have to ask ourselves how we can improve this today.
You describe Microsoft as an ethically-minded AI business. What do you concretely mean by that?
In the past, our thoughts focused above all on security, quality and software engineering. With all the progress we have made regarding AI, we have however realised that the quality of the software can still be improved. And we will have to work on this before we are able to consider ethics. But I want our company to focus on what it can control and do, how it makes the right decisions and the right investments and does not simply create any kind of AI – but rather an AI that caters to all the issues mentioned above.
As you touched upon earlier, artificial intelligence has penetrated more areas of society than might seem apparent at first glance. To what extent should a technology company ramp up its political responsibility?
The challenge is that technology has become far more mainstream than we could have assumed 40 years ago. But should we be following the same rules that other mainstream industries and technologies are subject to? As a multinational company active in 190 different countries, one of the most fundamental topics for us is ensuring that the contributions we make benefit the entire population. There is an economist whose work I have extensively read who examined how countries grow and the correlation between that and technology and its dissemination. The fact that we disseminate technology faster today than we colonised countries in the past is absolutely fascinating. Nevertheless, the challenge remains as to how intensively new technologies are utilised and what determines whether certain countries are more or less successful. If you are striving towards stable business relationships in the long term, then it is precisely this that is important. I, for example, do this not only because I want to be a good person, but also because capitalism has taught us that. If we think long-term, we will also generate long-term capital well into the future, which ideally will benefit many people and result in comprehensive economic growth. I think that this is where the economic and political responsibility for companies such as ours lies.
Satya Nadella is Chief Executive Officer of Microsoft. Before being named CEO in February 2014, he held leadership roles in both enterprise and consumer businesses across the company. Joining Microsoft in 1992, he quickly became known as a leader who could span a breadth of technologies and businesses to transform some of Microsoft’s biggest product offerings.
Picture credits © Microsoft