Akéo Plus and CEA Grenoble presented a research project at AI Paris which should allow artificial intelligences controlling robots to share data via a blockchain. A first proof-of-concept should be launched at the end of the year with the help of major French manufacturers.
Many start-ups are now seeking to bring together artificial intelligence algorithms and blockchain, the idea being to give visibility and traceability to the decisions made by the AI, whether for legal reasons or contractual. Specialist of Smart Robots, that is to say robots coupled to AI, Akéo Plus works with the CEA Grenoble to clear the concepts of what will be a blockchain dedicated to the industrial robots of tomorrow. “The challenge of the fourth industrial revolution is that everything will be connected and industry will be confronted with the problem of data decentralization. Smart robots will be more and more independent with robots that will no longer be content to repeat a gesture. The idea is to combine all these components of production, the edge computing and the blockchain. “, Highlights Stéphane Morel, founder of Akéo Plus.
If we extend the problem to the extended supply chain, multiple internal and external actors will generate data, which will increasingly raise the problem of the origin of the data on which the AI will make decisions, this data being able to be corrupted, modified and subject to debate between contractors if a problem occurs in the production line. “If tomorrow, in an agribusiness company, there is a failure on the production line and there is an impact in terms of public health, the challenge to justice, insurers and all the business ecosystem to find out who is guilty of what. Our goal is to be able to consult a register, an information drawn even if it comes from an AI. No industry today can guarantee its data. “Continues Stéphane Morel.
Data exchanged between robots of a production line, but also between factories
A 3-year research program will test industrial concept evidence to conceptualize and validate the approach to coupling AI and blockchain in the field of robotics. This program is conducted with the research team of Christine Hennebert, head of research and development of distributed systems and blockchain CEA Grenoble. The idea is not only to allow AI to write data in a blockchain but to design a platform at several levels: “We represent the system pyramidally, with the Smart Robot and its sensors at the bottom. Its signals are processed by a low-level AI and will aggregate the data to the benefit of the upper layer, the Edge Computing. This will allow sharing this information with other robots in the same assembly line. Finally, above that, we will share information between production sites, with other factories around the world. This entire chain of AIs must be secure, both horizontally, that is, sensor data, at the Edge, or vertically, when transferring data from one level to another. ”
Sensor data is stored in a blockchain so that it can be traced back to AI decisions. Christine Hennebert chose Ethereum to deploy the first proofs of concept: “We chose Ethereum for the moment for a very simple reason: it is the most advanced Open Source solution in terms of development tools. This choice allows us to move faster but what we are developing today is intended to be portable. The other advantage of Ethereum is economical. It is easy to set up a private blockchain with this solution while relying on a public blockchain would be much more expensive, the “Gas” Ethereum, the cost of the transaction Ethereum, being very expensive. The researcher adds: “The market is not mature and there is no blockchain” business “that would be directly usable by manufacturers. What interests us today is to advance on the concepts, on the high layers while keeping the possibility to interoperate with other blockchains. ”
Smart contracts within the supply chain
In addition to this data traceability, another capacity of Ethereum will be implemented as part of this research, the Smart Contracts. “We are going to create Smart Contracts in order to automate and streamline the passage of contracts within a Supply Chain, especially in the maintenance part,” explains Stéphane Morel. “Today managing these maintenance contracts remains very administrative, requires many exchanges of