The Case For Blockchain-Backed AI Capabilities

This might be an interesting way to earn some digital income: Become a part of Skynet and help spread the gift of artificial intelligence (AI) to in-need businesses around the world, all with the help of blockchain technology.

Yes, you read that right — Skynet.

But this is not the self-aware, AI-powered “creature” that nukes the humans in the “Terminator” film series before exploiting time travel to make sure its rather anti-social policies hold. This is the Skynet Project from a U.S.-based company called DeepBrain Chain, whose Chief AI Officer, Dongyan Wang, recently assured PYMNTS that its plans are designed to make it easier for other organizations to afford AI capabilities that could lead to better products related to payments and commerce.

The company aims to decentralize AI enterprise computing, harnessing the power of idle computers and building a blockchain-based offering that, according to DeepBrain Chain, can lower costs for AI services by up to 70 percent. The main problem the company wants to solve is “the expensive computing power and massive manpower to do the training of AI,” Wang told Webster during their conversation.

The Keys of AI

The proposition revolves around the realty that computer power, algorithms and models are all key to AI development — and that the high-level systems upon which online gaming, deep learning or other such activities are often idle at certain, even predictable, parts of the day. Why not use tokens to incentivize the use of those machines during slow periods to build a blockchain-based system that can offer secure, and private, AI capabilities to firms that have no financial incentive to invest in their own systems?

“It’s very expensive to do that,” he said, “and does not make economic sense” for many organizations, especially small- and medium-sized enterprises. Traditional cloud computing services might be too pricey for such firms, he added. As well, typical CPUs and GPUs in personal computers cannot rise to the level of AI computing. “A lot of AI companies cannot afford or do not want to have their own infrastructure in-house.”

In a way, Wang said, that pitch resembles allowing someone to rent your high-performing vehicle — or use it to make money via ridesharing — instead of just letting it stand cold in a parking lot. Or, think of it this way, he said: People who power their homes with solar panels can sell that power to utility companies when there is excess solar energy for their own uses.

The Blockchain Question

But will the companies that DeepBrain Chain wants to take on as clients be comfortable with using blockchain for the obviously important, costly and often sensitive tasks that require AI capabilities? Wang said the situation resembles the birth of cloud computing services from Amazon and other providers.

“A lot of companies said they would never put their data there,” he said. “But there were companies that were early adopters,” and now those cloud computing services are a significant, growing part of the overall digital economy.

“AI is very much in that space,” Wang said.

DeepBrain Chain stopped accepting candidates for its Skynet Project earlier this month. The company plans to launch a “TestNet” soon. A “MainNet” will go live in the second half of the year, according to the company’s plans, and its employees will continue to “research the possibility of large-scale AI distributed training and the whole [network-wide] distributed training,” according to a recent press release.

He was pretty mum on his company’s clients, but one does not have to think too hard to see how this could lead to AI developments and products in everything from connected cars to web-connected appliances — consider a refrigerator that gets even “smarter” via artificial intelligence and can basically do the grocery shopping and meal planning for you. Energy consumption could also be made more efficient via AI, Wang said.

There is little doubt that the money and research being devoted to AI will lead to significant developments in the short- and medium-term. A decentralized process to get AI where it is needed most, and at the right times, could help encourage that push.