Silicon Valley was yesterday, now comes the Silicon Economy," says Prof. Dr. Dr. h. c. Michael ten Hompel, Managing Director of the Fraunhofer IML. The industrialization of artificial intelligence is leading to the birth of a new Silicon Economy. It will become the battlefield of artificial intelligence.
The fact is that, driven by the huge growth in computer and sensor performance, storage capacity and broadband, real-time networking, the technological basis for this has long since been laid. "For the first time in history, we have significantly more technical possibilities in our hands, both in total and individually, than we can use sensibly," says the Dortmund logistics researcher. At the same time, a look at history shows that once technologies are in the world, they are also used. "Such disruptive developments are irreversible. Just as we can no longer imagine a time without smartphones, computers or steam engines, we will never return to the analog world of today after the introduction of artificial intelligence. The question is no longer if, but only when and how it will happen."

Logistics of all things

Where it will take place also seems to be clear: "The dawn of industrial artificial intelligence will take place in logistics," ten Hompel is certain. Less than ten years ago, his call for logistics to take a more leading role in the development of new technologies and conduct research triggered heated fundamental debates as to whether a service industry could conduct research at all. Why should the rather conservative logistics sector be one of the first industries to use AI processes? "Logistics is completely algorithmizable, it's in its genes, so to speak," answers ten Hompel. The individual process steps of physical logistics - getting the right goods to the right place at the right time - are simple and often even standardized. However, linking, networking and optimizing them is one of the most complex challenges of all - an ideal playing field for artificial intelligence. In addition, the highest scaling is expected in logistics. Logistics connects locations and companies in global networks. Practically the entire value chain is linked to logistics in one way or another - it is the basis for free global trade.

Artificial intelligence as a separate identity

It is also obvious, the researcher predicts, that large parts of the coming platform economy will be determined by its effectiveness and efficiency. Highly distributed and global artificial intelligences would emerge that learn and operate far beyond the imitation of human action. Intelligent load carriers and swarms of shuttles or autonomous, driverless transport systems are already the first evidence of this vision, and the mass application of intelligent sensors, smart pallet pools or autonomous robots is not far off.
But what should the algorithms that connect millions of devices in a future Internet of Things look like and what forms of artificial intelligence will reasonably emerge from this? Attempts to take nature as a model and transfer it to artificial intelligence are currently experiencing a renaissance. The ant algorithm, for example, is famous and can also be found in the first industrial swarms of autonomous vehicles. Taking nature as a model for biointelligent methods and processes in industrial applications is often due to the survivability of biological systems. They have survived for billions of years and have been able to adapt to the most adverse conditions over time. "With the digitalization of everything and artificial intelligence in everything, environments are now emerging that are not unlike a swarm," says ten Hompel. The focus of development is therefore increasingly turning to highly distributed artificial intelligence - especially for logistics. If we take this analogy seriously, he states, we will probably have to say goodbye to the classic, central ERP system in the long term, as there is no analogy in the nature of swarms.

The big picture of the Silicon Economy

Ten Hompel's "Big Picture" shows the living space and at the same time the playing field of artificial intelligence of a future, "virtualized" logistics in the midst of a data and platform economy. This begins with artificial intelligence and ranges from global networking and data sovereignty in a secure data space to financial flows via smart contracting and distributed ledger technologies. Using logistics as an example, it clearly shows the role that IoT, blockchain technologies and platforms already play in digitalization today. It also covers the entire data chain and shows how data will be generated, exchanged, refined and traded in a future data economy: from data generation by IoT brokers to the trading and booking of data by blockchain brokers and the organization of logistics processes using logistics brokers. The superimposed platforms enable new business models for a future data economy by combining data and refining it using AI and big data algorithms.

The basis for all of this is data. Artificial intelligence feeds on data. They need large amounts of data in order to learn in a meaningful way. According to a Gartner forecast, one in ten companies will earn money by exploiting their data as early as 2020. At the same time, over 20 billion devices are expected to supply this data via the internet by the end of 2020.

"The better the data is described and the more it is exchanged and processed, the more valuable it is. We therefore need to answer the question of how this data gold can be promoted in real time and handled in a sovereign and autonomous manner," ten Hompel sums up. At the heart of the Big Picture is therefore a secure data space that enables data to be exchanged and used in a self-determined manner. It forms the core, so to speak, of a - and this is crucial - federal structure of the data economy. The "International Data Spaces IDS" initiative already provides an answer to this by linking the use of data with the data itself for the first time. This allows data to be securely exchanged and used - because most data only becomes valuable when it is exchanged without being given away. His Dortmund colleague Prof. Dr. Boris Otto, Managing Director of the Fraunhofer Institute for Software and Systems Engineering ISST in Dortmund, is in charge of the implementation. According to ten Hompel, it is only logical that international data spaces in combination with new classes of (data) brokers must be a suitable means of networking and an integral part of the "big picture".

The scientist is convinced that the most important sources of data in the near future will be cyber-physical systems (CPS). From the doorbell, which uses a camera to recognize who is at the front door, to intelligent pallets that register when they are moved, to supermarket shelves that give personal shopping tips. Almost everything is to be digitized and provide individualized data. This will be made possible by the exponential development of computing power, which is not only taking place in the computers under our desks, but also in the embedded, cyber-physical systems of an Internet of Things. Ultra-low-power processors for less than one euro calculate with clock frequencies in the giga-Hertz range and provide sufficient computing power on site for a short time.

However, in order to trade data, it must also be transmitted. Fifth-generation wireless networks (5G) can ensure fast and low-power transmission, enabling battery life of up to ten years. Corporations are already relying on these new networks across the board. Others, such as the European Pallet Pool EPAL, are gaining initial experience with the 5G-compatible "Narrow Band Internet of Things NBIoT" from Deutsche Telekom, which also offers the appropriate flat rate via its partner company Once. Other networks that are not 5G-compatible, such as Lora and Sigfox, offer similar conditions and, in total, are expected to connect several billion IoT devices every twelve months over the next few years.

ten Hompel sees another effect of real-time networking based on 5G wireless networks: virtualization. "Cyber-physical systems now not only have significant computing power, but will also be able to access the unlimited power of the clouds and platforms to which they are connected virtually anywhere in the future. Distributed networks of artificial intelligence are emerging in a new dimension."

The final, decisive step towards the fully autonomous connection of material, information and financial flows in a
financial flows in a Silicon Economy are smart contracts. They are the missing link, so to speak, and ten Hompel is working on their implementation together with Professor Michael Henke, Head of the Fraunhofer Institute for Material Flow and Logistics IML.
at the Fraunhofer Institute for Material Flow and Logistics IML.
Cyber-physical systems, equipped with the appropriate software, are beginning to act autonomously. Henke and ten Hompel mean "act" quite literally here: contracts are automated according to defined rules, negotiated and concluded by software. This "smart contracting" in turn requires audit-proof booking and payment.

According to the current state of the art, both could be ensured in a distributed ledger and in distributed blockchain databases. Blockchain is the best-known distributed ledger technology and is often used as a synonym. It is an encrypted, tamper-proof and decentralized database and the perfect storage location for values, identities, agreements, property rights or proof of entitlement. The entries (blocks), such as a created Bitcoin, are distributed among the participants and cannot be subsequently manipulated. Everything that has been entered once remains in the blockchain forever. "This makes this technology ideal not only for cryptocurrencies, but also for the chronological and automatic execution of processes," says ten Hompel. For example, temperature data can be stored in a blockchain via a sensor attached to a refrigerated container if a limit value is exceeded. If damage occurs due to temperature differences, the chain of evidence is documented transparently and audit-proof via the blockchain and can in turn be evaluated automatically.

The "big picture" may be a dream of the future, but it doesn't just show the future, but what is currently possible. Behind every pictogram is a technology or a program that has been implemented and the majority of the software components are freely available as open source, at least in one version. "We have everything in our hands to make a total platform economy a reality - and not just in logistics," ten Hompel reiterates.

However, the "big picture" also shows an environment for artificial intelligence. It starts on a small scale, for example with image analysis, which allows a cyber-physical shelf to use a camera to recognize who is standing in front of it, also in order to give individual purchase recommendations or logistical instructions. It continues with autonomous vehicles that distribute goods and analyze their surroundings. Algorithms will also talk to customers and take orders, negotiate contracts and make payments. Ultimately, AI algorithms will permeate all levels and at the same time dissolve the levels themselves by means of virtualization. According to ten Hompel, the level models we are familiar with will only be a vehicle for the human imagination. "I am curious to see who will be the first to take this step consistently and set the de facto standards," says ten Hompel.

The race for the leading platform

Will the first one win the race in the new platform economy? The chances are good, says the researcher. Ten Hompel assumes that logistics platforms for the storage and movement of things and professional B2B platforms for large parts of supply chain management will soon follow. This industrial form of platform economy will primarily connect companies with each other in such a way that fully automated processing of business activities becomes possible. Ten Hompel has good news and bad news. The good news first: while the B2C platform market is already dominated by US companies, the market shares of logistics and industrial B2B platforms are not yet distributed.

Germany adorns itself with the title of "Logistics World Champion" as determined by the World Bank's Logistics Performance Index. Nevertheless, ten Hompel shares the fear - and this is the bad news - that American or Chinese companies will once again outstrip Europe unless massive investment is made in the development of adequate platforms and algorithms - from shelves to logistics cryptocurrency. This race will only be won by those who are prepared to open up and participate in neutral platforms. "Anyone who tries to make the world happy as an established company is subject to a late capitalist romanticism of bygone times and will soon only be driving parcels around for others," says the researcher, leaving no doubt about the future of the doubters.

But even the best platform with the most intelligent algorithms will not succeed in the new market on its own. Uber and co. have bought the markets with disruptive offers and marketing worth millions - and they will do so again, ten Hompel is convinced. But the Europeans have also learned their lesson and will no longer be taken by surprise so easily, he believes. "One percent of turnover for logistics research" was ten Hompel's call to logistics service providers seven years ago. Dismissed at the time as a pipe dream, today some logistics service providers have already exceeded the target.

The challenges of digitalization are by no means only of a technical or economic nature. "Of course, human workers will be replaced by machines and their artificial intelligence," says ten Hompel. This is as unstoppable as the introduction of the steam engine or the computer. His conclusion: "We will neither be able to predict nor determine the future, but we can shape it. Only if we succeed in playing a significant role in an industrial platform economy with its distributed artificial intelligence will it follow our social norms and goals. One thing is clear: the world will turn upside down with and without us - because it can."

About Michael ten Hompel

The "logistics pope" is a professor at the Technical University of Dortmund, where he holds the Chair of Materials Handling and Warehousing and is Managing Director of the Fraunhofer Institute for Material Flow and Logistics. ten Hompel was the cover hero of the first issue of LT-manager, which was published in October 2010. In between, the Dortmund native has made many correct predictions, been inducted into the Logistics Hall Of Fame (2012), been part of the highly successful LogistikRuhr efficiency cluster and engaged in many exciting discussions about the logistics of the future. In February 2019, it is now time to present his next big hit - exclusively in LT-manager, of course.