Joseph Sirosh, Head of AI and Corporate Vice president of artificial intelligence at Microsoft International, talks very openly about the realities of AI and why it’s not about replacing people with machines. What Microsoft does to make sure that AI serves people and how Thyssen and STRABAG already use AI in Austria is revealed in this interview.
Leading economic experts see huge potential in the use of artificial intelligence (AI), but we also repeatedly hear voices painting bleak end-of-the-world scenarios. What are the prospects of this technology, in your view?
Every ten years or so, another trailblazing new innovation comes along that revolutionizes the digital landscape. In the 1990s, it was PCs; at the turn of the millennium, it was the internet boom; and ten years after that it was mobile devices. All these new developments brought with them an undreamed-of potential for companies that were bold enough to adapt. The latest innovation in our increasingly digital world is artificial intelligence.
When we talk about artificial intelligence, what we actually mean is complex algorithms that search for recurring patterns in huge mountains of data in order to find a solution to the problem at hand – instead of simply responding to certain input instructions.
Microsoft has developed a number of exciting prototypes that illustrate just what the technology has to offer. Our intelligent cane automatically calls for help when a visually impaired person trips over, an intelligent glove interprets sign language and our Seeing AI app gives blind people a description of their immediate environment.
Robots are faster, more efficient and more rigorous than people. Should we be scared about robots stealing our jobs?
It is particularly important for us to emphasize that the goal must never be to replace people with artificial intelligence. Quite the opposite, in fact: people and machines complement one another very well and make an unbeatable team. Our empathy, creativity, emotionality and innovative power combined with the algorithmic abilities of artificial intelligence give us the capacity to do great things in the future. We absolutely should not be scared about a real robot revolution!
Pioneers in AI are demanding specific regulations and standards with regard to the development and application of AI technologies. Do you consider this approach a meaningful one, or do you see it more as a hindrance to the innovative power of the technology?
It’s true that artificial intelligence is exercising increasing influence over our everyday lives. Microsoft has been investing in artificial intelligence for almost 25 years now, and aims to offer private customers, businesses, and developers the widest possible range of options that make the most of this promising technology. In order to mitigate people’s fears about AI and make the decisions it makes more transparent, specific regulations and standards are indeed essential. Microsoft has established a total of six binding design principles to this end:
- “AI must assist humanity”: Industrial robots can take over dangerous, monotonous or physically demanding tasks, thus taking the strain off people working in these areas.
- “AI must be transparent”: People must always be able to understand which systems of rules AI decisions are based on.
- “AI must maximize efficiencies without destroying the dignity of people”: Artificial intelligence must support people without infringing their human dignity.
- “AI must guard against bias”: Scientists and developers must always ensure that AI applications make impartial decisions and that there is no chance of misuse.
- “AI must be designed for intelligent privacy”: AI technologies must be designed in such a way as to ensure people’s privacy permanently.
- “AI must have algorithmic accountability”: Uncontrolled use of AI must not be permitted and there must be no chance of unintentional malfunctions.
The advantages offered by artificial intelligence do sound very positive. Is the technology already being used by companies, or is it a Silicon Valley trend that will only become relevant to us in a couple of years?
A current poll* of over 800 executives from global companies has revealed that AI technologies are already in use around the world and in all sorts of fields at that: monotonous routine activities are seeing increased automation, human teams are receiving ongoing AI support in the form of comprehensive metadata analyses, and business risks and challenges are being evaluated faster and more accurately than ever before, thanks to AI.
Digital transformation and sophisticated Industry 4.0 environments are making companies more agile and creating huge revenue opportunities. Thanks to innovative technologies such as artificial intelligence, mixed reality and the Internet of Things, we are able to tackle the biggest challenges of our time, such as water shortage, poverty and disease, and the aging population.
The digital economic miracle, however, has long ceased to be merely a vision of the future: Austria’s decision-makers must start setting the right course today so that Austrian businesses can remain competitive on global economic markets in the future too.
*Tata Consultancy Services Study: Artificial intelligence is a top priority among companies (April 2017)
Can you give us a specific example of AI from Austria?
The industrial group ThyssenKrupp is putting great emphasis on Microsoft’s AI technologies in order to connect its elevators. This means that all the functions and data of the elevators can be viewed in real time, and they can anticipate problems and maintenance measures. STRABAG, a leading technology corporation for building services, is also looking to artificial intelligence as part of a large-scale infrastructure project. Selected raw materials storage will be automated, constantly monitored, logged, and a variety of predictive maintenance measures implemented.
AI-based applications draw their intelligence by analyzing huge quantities of data. How can you reconcile this innovative technology with the issue of data protection in a meaningful way?
Artificial intelligence is based on mathematically extremely complex processes, which are sometimes difficult to fully comprehend at first. The decision-making process of an AI application is therefore frequently described as a black box. Microsoft is fully aware of this and therefore committed to ensuring that we do not supply developers with an unfathomable complete solution, but rather a variety of AI components, for example in the fields of image and language recognition, so that our customers can develop their own AI solutions.
Admittedly, defining the framework conditions for the application of artificial intelligence is by no means a trivial matter. Microsoft, however, has comprehensive expertise in data protection as well as in maintaining privacy. Furthermore, Microsoft has undertaken to adhere to the EU-US Privacy Shield as well as any possible EU standard contractual clauses. We also believe that the EU Data Protection Basic Regulation is a key step in clarifying and enshrining individuals’ data protection rights.
Even though it is apparent that the framework conditions for the use of AI applications cannot be clarified definitively today, politics, business and society still need to put their heads together on an ongoing basis to address the question. This is the only way to ensure that the diverse opportunities offered to us by AI technologies are truly sustainable.