The Value Foundation of AI and Challenges
The value foundation of AI rests on the raw materials necessary for producing it, namely data and computing power, while real social and economic value is created by AI algorithms. Algorithm models are the layer that’s closest to us, as AI applications are only made possible by them. In other words, the value of data and computing power is actualized through algorithm models. Without the models, any amount of computing power is futile.
In the past few years, cheap and efficient computing power (made possible by GPU), a near-infinite storage capacity for big data, an omnipresent deployment of sensors and the fast development of all other supporting technologies have created the explosive growth of AI. And with this comes the biggest challenge: how to protect the intellectual property of algorithms, thus protecting the value foundation of AI so that AI scientists are motivated to continue creating.
Protecting AI intellectual property is challenging in the following ways:
Difficult to authenticate:
Like with other software products, it is difficult to authenticate the ownership of an AI algorithm, which subjects the works of AI scientists to theft and abuse.
Easy to copy:
Although costing a lot to create, an algorithm is essentially a string of code, and it takes no more than a few clicks to copy it. Being no experts in encryption techniques, most AI scientists lack the effective means to protect their intellectual property.
Difficult to protect:
In a centralized environment, anyone who has access to the system can easily copy other people’s algorithm models for free.
These challenges will shake the value foundation of AI in the long run and even strangle its growth potential. Fortunately, with the help of blockchain and distributed storage technologies, we are well-equipped to tackle these challenges.