As you probably know, startups that combine Blockchain and artificial intelligence technologies are all the rage now. For example, the SingularityNET project, which intends to create a Blockchain ecosystem to host a lot of specialized AI systems, has raised more than $150 mln from almost 10,000 investors via an ICO. Other Blockchain projects that use artificial intelligence raise substantial funding, too.
The rising popularity of such startups is understandable: AI is the future! One of the most promising application areas for artificial intelligence is the healthcare industry, which is why more and more Blockchain-based healthcare AI projects appear. The million dollar question is, is any of them going to succeed?
When evaluating the chances of success for such projects, you need to keep in mind the core idea of artificial intelligence: any complex AI must be trained, otherwise, it’s mostly useless. In case of a healthcare AI system, you need to provide massive amounts of data (“big data”), so that it can learn. The more data you provide, the more accurate the AI becomes.
Big data issue
The problem is healthcare big data is extremely expensive. Small, independent neural network developers can only dream about having it. Even corporate giants can scarcely afford a sufficient amount of such data! For example, IBM has obtained only about one percent of the big data it needed for training its famous AI system, Watson, by acquiring a leading provider of cloud-based healthcare data and analytics for $2.6 bln.
As you can see, the development of AI for the healthcare industry is hindered by a lack of healthcare data marketplace. But do the Blockchain-based healthcare projects address that problem? The doc.ai startup, which has recently raised about $10 mln via an ICO, intends to obtain healthcare data from the end users (patients) by providing medical advice in exchange for their medical histories. Generally speaking, only the patients who need medical advice will be interested in uploading their healthcare data to such systems. As the doc.ai NEURON brochure says, “[i]n the future [people] will even be able to monetize their health data on a monthly basis by leasing it to pharma companies.” It doesn’t look like the doc.ai project is going to create a large-scale healthcare data marketplace any time soon.
Golem and SONM are Blockchain projects that aren’t aiming at the AI market in the healthcare sector yet, but they possibly can host healthcare AIs on their decentralized platforms. However, if Golem or SONM try to do that, they are going to encounter the problem we’ve already mentioned: a lack of healthcare data marketplace. It means none of them will be able to train their neural networks.
Another potential player in that market is SingularityNET, which provides an infrastructure for AI hosting and internetworking. Unfortunately, it doesn’t address the problem of a lack of access to healthcare big data. According to its whitepaper, SingularityNET intends to become a dominant provider of AI services in different industries, including healthcare. We feel sorry for SingularityNET investors, but that risks not happening because Skychain is going to grab the dominant share of the healthcare AI market!
So far, Skychain is the only project based on a clear, actually working approach to providing big data for neural network training.
Skychain will provide an infrastructure to host independent neural networks, which will be trained on massive amounts of datasets (medical histories from patient cards, medical reference data and medical research data) uploaded to the Skychain system by data providers. As a result, the neural networks hosted by Skychain will make the most accurate diagnoses and prescribe the most suitable medical treatment.
You may ask, but how is Skychain going to create a healthcare data marketplace? The answer is, Skychain will offer an economic incentive. Healthcare data providers will be getting a reward for each use of any neural network trained on their datasets. As a result, the amount of big data available for neural network training will grow exponentially, and the AIs hosted by Skychain will soon become smarter and even more accurate.
Many healthcare AI projects are going to fail as they cannot motivate healthcare big data providers to share their datasets for neural network training. None of the existing Blockchain projects for the healthcare industry seems to offer a viable model for creating a healthcare data marketplace, so Skychain could be the only hope of high-tech medicine.
The Skychain project is going to pre-sale its tokens with a 50 percent discount from Dec. 18, 2017 to Jan. 7, 2018. Prior to the ICO, Skychain representatives will speak at eight or more major crypto conferences all over the world.
Most importantly, the Skychain project is based on existing approaches and technologies, and all the legal and regulatory issues have been worked out. So hopefully all Skychain token buyers will make a nice profit when the project is implemented.