In the coming years, the use of Artificial Intelligence applications will grow exponentially, and the use of these capabilities will be extended to all fields. Currently, there are still some barriers to make this growth feasible. For a few years now, the Blockchain platform has been betting on one of the accelerators to multiply the supply and boost the demand for AI capabilities in a generalized manner.
As the study The Impact of Artificial Intelligence on Entrepreneurship in Latam conducted by everis and Endeavor indicates, the main difficulties for AI growth in Latin America are:
- Shortage of AI specialized talent, the labor market lacks professionals with the kind of technical knowledge required to develop these solutions.
- Lack of data to train, the most advanced techniques of Machine Learning require large volumes of information to achieve robust models with high precision.
- Lack of Knowledge regarding the value or income that AI solutions can provide, customers are not familiar with this type of solution, generating some distrust regarding the products and services proposed.
- Difficulty in finding funds that have as objective to develop AI capabilities.
Blockchain is a technology devised by Satoshi Nakamoto in 2008 as the platform that would support Bitcoin, a digital currency characterized by being highly secure, with decentralized management, distributed platform, without intermediation and with finite offer. Due to its properties, the Blockchain has been placed as the ideal platform to materialize the Smart Contracts, an idea postulated by Nick Szabo in 1994. As a consequence, the Blockchain is one of the most promising technologies and a great receiver of investments.
Blockchain can revolutionize a large set of industries, for example:
- Energy: democratizing energy trading.
- Food: giving information to consumers about traceability and certifying food.
- Medicine: allowing the creation of a universal, centralized and secure base with the medical records of patients.
Artificial Intelligence in the Blockchain ecosystem
To the existing initiatives that are working with AI and Blockchain, I divide them into two types: Competitive Platforms and AI Large Storages, detailed below:
There are several projects in the startup format that are working with Blockchain technology to create competitive, AI, decentralized platforms. For example, the Algorithmia Danku platform developed over Ethereum, a platform that has as an objective similar to Kaggle.
Kaggle, founded in 2010 and purchased by Google in March 2017, is the largest community of data scientists, with over one million participants. The main objective is to serve as a platform to perform machine learning competencies in which an organization or company publishes a challenge, the competitor with the best result wins an economic reward. This continues to be the core of the platform, although it also performs educational platform functions, a public database repository and even a workbench for data science.
Danku, unlike Kaggle, is built on Blockchain technology, and the competences are published in smart contract format and is constituted by the data, the evaluation function and the economic reward. Competitors submit their neural network, the platform performs its analysis using the evaluation function and finally sends the reward to the competitor with the best model.
Competitive platforms are a powerful alternative for acquiring talent provided that the use case is well defined and is determinant for the case to get a model with high precision. These marketplaces of talent will end up developing a global market for professionals and companies of Deep Learning, better serving future demand and promoting the creation of better AI applications.
AI Department Stores
There is another type of initiative that organizes decentralized marketplaces around AI, the objective of the platforms is heterogeneous, with objectives such as sharing information, sharing computing capacity or integrating ML models within the Blockchain. The most ambitious ones propose to organize large ecosystems in which the participants of the AI community can find what they are looking for:
- Data providers. Thanks to its level of encryption, Blockchain is perfect for storing sensitive information as a support for individuals and public entities to exchange information. The use of Smart Contracts and the execution of Machine Learning models within Blockchain ensure that the data is used by the consumer in the agreed manner.
- AI services consumers and algorithms providers. The use of Blockchain as a technological base will mitigate risks for participants. The execution of the models within the Blockchain guarantees the privacy of the data provided by the consumer and also reinforces the privacy of the model, along with the use of smart contracts for financial settlement.
There are experiences on Ethereum that allow to train and infer Machine Learning models within the platform using the main libraries such as TensorFlow, Pytorch, Caffe, etc. These models could be used from outside the platform or transparently by active smart contracts in the ecosystem itself.
- Suppliers and Consumers of computing capacity for AI. AI is a large consumer of computational capacity. Therefore, for the development of AI it is essential to have the necessary computational capacity at all times for a reasonable cost. Blockchain allows suppliers to offer excess capacity to the marketplace, providing a guarantee to consumers of security and privacy, associating via Smart Contract and digital currency the payment for use.
There is experience on Ethereum with this format that allows to use not only CPU but also GPU.
- Data Consumers, in many cases the barrier to reach the state of the art of a model or viable AI application is not having access to the volume of information needed.
Frequently out of zeal, compliance and data protection regulations make it difficult for companies or public bodies to be willing to share information openly. The use of Blockchain platforms and the evolution of the Homomorphic Encryption will allow the models to be trained in a secure way having access to the encrypted information, which the model will benefit from large databases without compromising the confidentiality of the information.
These AI ecosystems will enhance the growth of AI. The strength of these platforms lies in the creation of a community that transmits the necessary trust for the aforementioned actors to participate, creating data offerings, facilitating the access of talent to the demand, creating safe Machine Learning as a service platform, multiplying the offer of computing capacity, etc.
There is a great variety of startups in different phases of evolution that are working with Blockchain technology with the aim of constituting ecosystems to offer AI services in Data such as Ocean Protocol, DX Network, SingularityNet, Deep Brain, Cortex, Consensus or Golem.
Returning to the study of the impact of Artificial Intelligence on Entrepreneurship in Latam (1) carried out by everis and Endeavor, the study indicates that the main challenges for the development of AI is the low Adoption of Technology and the low level of Investment that they indicated as the first challenge, respectively, by 39% and 33% of the interviewees. The biggest barrier to the adoption of AI is to demonstrate the value this technology brings to business and teaches the correct way to use it to obtain good results; once these barriers are overcome, investments will naturally arrive.
The AI platforms on Blockchain will be very important for the AI expansion, although for large business world communities to gather around these platforms, organizations will be required to learn how to use AI.