Blockchain Can't Handle Medical Big Data? This New Technology Says 'No!'

The value of AI to the healthcare industry is undeniable. Currently, acquiring data sources has become the primary bottleneck for AI. The emergence of blockchain technology has brought a ray of hope to the industry. Essentially, unlike AI, which focuses on breakthroughs in “productive forces,” blockchain technology primarily addresses issues related to “production relations.” Of course, blockchain technology also has its unique advantages in enhancing productive forces.

Author: Jiang Jiang, Founder of Chengdu Yingda

If the medical AI industry merely relies on rigid, outdated methods to acquire data, its conservatism will impose higher costs on enterprises.

We often hear stories of industry giants occupying resource high grounds, building their own moats, and ultimately forming monopolies to sit securely in a dominant position. Blockchain, on the contrary, is akin to "Tai Chi" martial arts; the process of acquiring data is by no means rigid, but rather employs "softness to overcome hardness." It is crucial for the medical AI industry that patients can control their own data. The core philosophy driving AI is to proactively embrace and establish patient data trading markets, rather than engaging in predatory acquisition.

Medical AI companies need to band together to establish a distributed data trading marketplace, reducing their operational costs while ensuring data sources receive fair compensation.

Blockchain excels at building ecosystems, as its incentive mechanisms encourage active and voluntary participation. The payers for healthcare services will shift from patients to commercial insurance companies. Medical AI companies must leverage blockchain to either build or proactively join such an ecosystem.

However, neither Bitcoin nor Ethereum was designed with large-scale data storage in mind. Although Ethereum features the concept of storage space, storing approximately 1 MB of data would incur prohibitively high costs for enterprises given current technological capabilities.

In 2016, the industry consistently avoided using blockchain for storing large files and big data, as no ideal solution existed. A common refrain was, “Keep core data on-chain, but store large files on traditional cloud storage.” However, this approach introduces new complications and costs, particularly concerning issues such as rights confirmation.

The era of “democratic healthcare” begins by empowering individuals to own their data. If large files are stored in traditional cloud storage, asymmetry between patients and medical institutions persists, and the risk of privacy breaches remains. Not all data needs to be recorded on a blockchain; however, high-value data, regardless of size, should be stored in a decentralized manner, with private keys theoretically held by the patients themselves.

This is the overarching trend in healthcare and the fundamental driver for all participants in medical innovation. Only by respecting individual privacy can a market economy thrive—a timeless principle.

Within blockchain technology, IPFS (InterPlanetary File System) is a relatively unique technology, translated as the "InterPlanetary File System." Timing is everything; IPFS technology happens to offer a solution to the problem of decentralized storage for big medical data.

If AI companies wish to build a global collaborative network, they would incur substantial costs in cloud storage. However, by leveraging IPFS technology, such a network can be established at a low cost. Essentially, IPFS enhances AI productivity. In summary, this is achieved through two key benefits: improved storage efficiency and reduced storage costs.

Distributed storage offers significant advantages over centralized HTTP-based approaches. With the HTTP protocol, users can download only one file at a time from a single server and cannot retrieve files simultaneously from multiple computers. By adopting IPFS’s peer-to-peer (P2P) transmission method, bandwidth costs can be reduced by 60%. Practical tests have shown that uploading a 500 MB file using an IPFS-based consortium blockchain storage architecture takes less than 10 seconds.

A blockchain operates on the IPFS network to store a table of hashes for internet files. Each time there is a network access, the address of the content (file) must be queried on the chain.

IPFS envisions a network where all terminal nodes do not merely act as browsers or clients; instead, everyone can serve as a network operator and a storage node. By integrating technologies from both CDN and BitTorrent, IPFS enables data retrieval from multiple nearby nodes.

Each node’s DHT (Distributed Hash Table) stores file segmentation information, hashes, and address details. However, this is not unique to IPFS; anyone who has used eMule should be familiar with DHTs (Distributed Hash Tables).

DHT has the following properties:

Discrete (Autonomy and Decentralization): The nodes constituting the system are peer-to-peer, with no central control mechanism for coordination;

Scalability: The system is required to operate efficiently regardless of the number of nodes.
Fault Tolerance: The continuous joining and leaving of nodes does not affect the overall operation of the system.

Leveraging these features, developers can utilize IPFS to build a distributed peer-to-peer network between AI companies and medical institutions. Uploaded file information is “sharded” and distributed across IPFS storage nodes. The storage network has no central authority; every authorized user can download files at any time. Moreover, similar to BitTorrent downloads, the more nodes there are, the faster the download speed.

Moreover, IPFS has developed an incentive model based on Filecoin. Filecoin can be understood as a token for cloud storage, where users pay tokens to providers who share their storage space. Leveraging this incentive network, IPFS ensures that valuable information is preserved permanently and immutably, preventing user data from being deleted due to the shutdown of services by any single company.

To maximize profits, AI startups must adopt a decentralized philosophy to attract more customers. Nebula Genomics, a startup founded by Harvard genetics luminary George Church, is one such company.

In George Church’s view, the dominant force in gene sequencing transactions will be individual customers, rather than sequencing companies. Users can acquire Nebula tokens through transactions to purchase sequencing-related services. Ceding data trading rights to customers may appear to strip enterprises of their “dominance,” but it actually fosters a virtuous business cycle. When patients are assured that their privacy is protected, they become more actively engaged in AI-driven genomic analysis services. This shift in the primary transacting party does not diminish the competitiveness of AI companies; on the contrary, it strengthens their persuasiveness when securing external partnerships.

In the fields of gene sequencing or related AI, the involvement of insurance companies offers hope for profitability. UnitedHealth, the largest health insurer in the United States, is also undertaking similar initiatives. Payment based on AI performance will serve as a key incentive for insurers to cover these costs. Due to the immutability and traceability of blockchain technology, users can track the therapeutic efficacy of AI systems and conduct benefit analyses based on diagnostic and treatment outcomes.

In 2018, blockchain moved beyond mere hype and entered its first year of practical application. Particularly for on-chaining large file data, IPFS is undoubtedly an excellent choice.

Reducing costs through blockchain technology represents only a small fraction of its value. More critically, it drives customers and payers to build an ecosystem via “decentralization,” encouraging customers and partners to contribute data more proactively and receive support in return.

Leveraging next-generation blockchain technologies such as IPFS, EOS, and IOTA, it is possible to build a “trustless” and “self-incentivizing” healthcare network. Within this new network of production relations, any individual can launch their own ambitious venture, significantly reducing transaction costs.

Blockchain not only powers AI, but it also gradually breaks down information silos in healthcare. By reconstructing production and trust relationships to reduce medical costs and improve efficiency, blockchain makes its greatest contribution to this era.

This article is a guest contribution and does not represent the views of VCBeat.