Crypto is decentralizing, AI is centralizing, according to Peter Thiel. Although the venture capitalist has followed this remark with a somewhat strange ideological classification for these technologies, the premise rings true. Artificial intelligence advancement is now in the hands of huge companies such as Google, IBM, Microsoft and their Chinese counterparts.
Machine learning relies on data – the more the better – and platforms such as these have proven skilled in collecting it. They have used their competitive edge to make AI products better which draws in more users and more data – a great example of leveraging network effects. However, this has also created a problem for small actors that want to get in on the AI game.
“To create AI applications, developers need to write algorithms and use machine and deep learning but they also need images or videos to send this raw data into the algorithm’s function. This enables them to train the machine learning system to get results,” Clement Duan, founder of AIChain, told TechNode. “But, as we know now, those data resources are controlled by big companies like Alibaba or Google. It’s really hard for small and medium-sized companies to obtain data.”
AIChain is one of the companies trying to democratize AI with blockchain. Similar to SingularityNET, the China-born Singapore-based platform is trying to connect public users with AI resources. Founded by Duan, a former director of software development at the world’s largest bitcoin miner Bitmain, the company has so far drawn investment from Bitmain, INBlockchain, Viking Capital, China Galaxy Securities, and RadarWin.
According to Duan, big companies that hoard data is not only a problem for SMEs, online users also have very little control over their data. Take shopping as an example: big companies use buying records without our knowledge to make money.
“We want to reach a stage where everyone can share their resources on the blockchain and open to others. The public, the small businesses, and individual developers will get a chance to obtain data and AI resources and create their own AI product and application software which will improve their efficiency.”
Of course, as recent data privacy scandals such as Facebook’s have shown, sharing data is not easy. Users need to be willing to share their digital assets like images and videos, algorithm modules, and applications such as software or tools. Naturally, companies will worry about losing information and data resources, including private information.
“Why do these people want to share their resources on the blockchain to others? We give them a chance to benefit from their resources,” said Duan.
AIChain plans to solve the privacy issue with blockchain and DRM (digital rights management) technology. To use data, which is encrypted, watermarked, and fingerprinted, users have to pay a digital token to the owner of a particular resource which helps verify the buyer through blockchain transaction records.
Marketplaces such as AIChain are not the only example of marrying blockchain and AI. Projects such as DeepBrain Chain are trying to achieve similar democratization of resources as AIChain but instead of algorithms and data, the platform is focusing on distributing computing power.
Another area is supply chain management which is increasingly relying on automation through AI to boost efficiency and reduce costs. Chinese e-commerce giants such as JD and Alibaba are currently trialing blockchain to perfect their huge supply chains. The AI blockchain combo also has a bright future in market analysis and forecasting.
However, Duan thinks that blockchain is not a panacea—its good for recording short but important messages, for opening information to the public, and it is hard to hack. However, the transaction efficiency of blockchain is still very low and the technology has no function in protecting data content. If we want to benefit from blockchain, we will need to apply it to existing projects, said Duan. One example is using smart contracts to cut costs and improve efficiency for platforms which can help them get more users from the public.
“The year 2017 saw too many blockchain projects. Many projects might be useless or may disappear in 1-2 years. Blockchain tech has its limits too, it cannot do everything.”