The ethics of AI: "Leave the human in the loop"

This is what Marisel Kehbog, the chief planning director at SAP, advised in a lecture on the ethics of planning at the 2018 SAPPHIRE conference ● She also said that "there is still a lot of work to be done to make artificial intelligence scientifically proven - that is, that it can be produced, measured and explained"

Keep the person in the loop, this is one of the tips given by Maricel Kehbog, chief planning officer at SAP, in a lecture on the ethics of planning at the SAPPHIRE 2018 conference, which took place recently in Orlando, Florida. But according to her, first of all, in order for us to be able to trust artificial intelligence, we must turn it from alchemy into a scientifically proven one - meaning that it can be produced using it, measured and its decision explained.

"Technology has always brought change to our society - ever since the industrial revolution. Even then, there was a disappearance of existing jobs due to new and better jobs. This routine has continued to this day," said Kehbog.

"Will it be different with the rise of artificial intelligence?" Kahabog asked. "Today there are many high-demand jobs, which were not in the past, as well as the high demands that were not in the past. In the future, new skills will be required for jobs that we do not know at all. What advice would you give your children to prepare them for the future market? Change is fast, and agility is talent The best you can do for yourself, and in addition, lifelong learning is the new normal," she stated.

"Everyone is afraid that AI will take over the existing jobs and push them all out. We can use technology to help us prepare to work together with machines in every business parameter. We at SAP believe that we must provide the tools so that users can take advantage of the capabilities of artificial intelligence without the risks.

With smart solutions based on machine learning and augmented reality, we can imagine a future where technology supports the development of work, helping us personally develop based on our skills, helping us learn more and work confidently in our work. It is good for us and the organization. This way we will also learn new skills. Gives access to better internal information and better data. Artificial intelligence that is built correctly can lead us on the right path today."

"The missing aspect is trust"
"There is still a lot of work to do to make artificial intelligence scientifically proven - that is, that can be produced with it, measured and explained," said Kehbog, "but even when it becomes scientifically proven, there is a missing aspect and that is trust. We have to create a relationship with AI based on About trust. We need to create AI systems that are consistent, reliable, and trustworthy."

"When someone works with smart artificial intelligence systems, they need to know that the data they receive from them is based on reliable information. An explanation in the system that will be able to detail to the user how it arrived at the information it displays, will help the user know that he is making the right decision. Also, if the system does not have information provides, it must know how to go look for the missing information and provide it, while it explains it to the user. The priority of the system should always be to help the user. If the system is not able to provide information to the user, it should find alternative ways to help him and direct it, such as using a different algorithm, so that the user can be sure that the result is correct, if it is the same in both algorithms."

The principles of building artificial intelligence according to Kehbog
Smart automation - "We have to be careful with what we make smart and we also have to do it smartly ourselves. We have to ask, will the machine be smarter than the human in this specific task? For example, can we imagine our parents in a hospital where there are no human caregivers but Only robots and machines? The point is that before we automate an action, we need to think about everyone involved in that action."

Human augmentation - "The way to combine human abilities and technology is through innovation. When people work together with artificial intelligence, we can achieve great achievements. People approach different tasks differently than AI approaches them and this is what will result in a winning combination."

Digital ethics - "We must and should build methods of digital ethics into the systems we build. We must allow artificial intelligence to learn and also to forget. We must train the systems not only on data but also on familiarity with the human point of view. And we must always provide an exit door through which A person can take over the systems when it is necessary."

Keeping humans in the loop - "All this can grow if we feel we have control. It's not science, it's alchemy. AI cannot be a black box that provides us with products without the possibility of explaining or reproducing them. We must know how to understand the machine, explain it and finally control it in it".

SAP's solutions on the subject
"One of the tools we are developing that allows us to receive internal information about information that we did not receive before in an experiential configuration is the SAP Inscribe - a new way to analyze your data. The solution combines the most human interfaces such as paper and pen with cross-border technological solutions and machine learning.

It enables quick decision-making, while being able to write notes in the form of a pen on documents and reports, for maximum convenience. The solution allows us to receive the data in the way we want to receive it, in a way that is convenient for us. The solution provides our business with critical analytics capabilities through the SAP Smart Business for SAP S/4HANA.

Now you can make cuts and analyze the data in a way you couldn't before. The SAP Inscribe solution will be available on SAP S/4HANA next year. And SAP Analytics cloud will be available after that."

In conclusion, Kahabog says: "For 46 years, SAP has been helping businesses in the transition to become large organizations. We have helped them go through digitization and navigate the path to the cloud. Today we are already working on the next challenge - to turn organizations into a smart enterprise.

Proper construction of artificial intelligence will help SAP prepare customers for an innovative and exciting journey. AI will be able to fill the required skills in the future. We believe that all customers should be made to accept artificial intelligence and trust it, and this can only happen if we build it correctly."