Deep Learning, Robots, and the automation of repetitive tasks in businesses are frequently the first things that come to mind when people hear the term “Artificial Intelligence.” Nowadays, Artificial Intelligence research is focused on specific areas, a practice known as narrow AI. This type of AI cannot yet draw conclusions based on common sense.
The difficulty of eliminating biases in the results generated by the algorithms used, which are trained on data that, in its original form, often contains these vices, is currently the main impediment to the widespread adoption of artificial intelligence at scale, and which ultimately results in the automation of decisions that have the potential to affect people’s lives.
People are generally unaware of these facts, and as new data protection laws are implemented around the world, the development and use of artificial intelligence in areas such as healthcare and government services that deal with sensitive data will necessitate the implementation of a Data Governance Strategy that includes human supervision in order to avoid bias and errors in the data.
To address this issue, it is necessary to make “clear” which premises and factors were considered in order to generate a result from the machine learning algorithm. To put it another way, the ability of a human operator to explain and interpret the results of this algorithm in relation to the context of the problem at hand.
Data privacy laws and the process of consent management
Also, there is the issue of who owns the data, and in many countries that adopted Data Privacy laws like the GDPR in Europe and LGPD in Brazil, the users own the data and they can choose to whom they are going to send their data in order to have a better and tailored user experience and improved products and services.
Visibility to users of the traceability of their data, once explicit consent (from the user) is given to organizations for a specific purpose, is essential for increasing public trust and transparency throughout the process. As a result, the use for a different purpose from the original, in our view, should require specific new user consent.
For example, organizations in many countries in order to be compliant with Data Privacy laws must implement a Consent Management System and processes with Data Governance policies to support it.
Also, once users revoke the permission given, the act of opt-out from a database of an organization, no longer their data can be used to train AI models. The implementation of this process can vary according to the specific Data Privacy laws of each country.
The increasing adoption at the scale of AI Applications worldwide
Once one application is adopted in a specific market or organization, others try to adapt the solution to its own reality in order to remain competitive.
In this so-called “technology adoption life cycle” process, the innovators and early adopters of technology tend to obtain a competitive advantage over potential competitors due to their pioneering spirit.
The result of the adoption of AI Applications at scale worldwide, in our view, would be the improvement of the decision-making process in the management of companies, societies, and governments in general, with gains in productivity that will result in lower prices for products and services, that may result in an improvement of Healthcare access and costs, also helping to solve poverty problems around the world.
World Market for Data Monetization – Major Share in North America
The data monetization market is expanding rapidly as a result of increased investments in Digital Offerings by major corporations such as Google LLC, Microsoft, Amazon, and IBM Corporation.
Also, much of the increase in data volume will come from M2M communications and IoT devices at the Edge of IoT networks across the Internet. In this case, all this generated data needs to be processed and analyzed in order to generate value in the form of actionable insights.
Generally, part of this data can be processed on the Edge, but a huge part is sent to be processed with the use of Advanced Analytics in a Cloud Computing Environment.
The trend for high-intensity computing in organizations to take place in a Hybrid Cloud is increasingly becoming a reality. Gone are the days when an On-Premise environment could cost-effectively meet all business needs.
For example, the average Internet user is very aware of Google Search Engine and services like Gmail. Nowadays, Google controls about 62 percent of the mobile browser market, 69 percent of the desktop browser market, and approximately 71 percent of mobile devices’ operating systems (including smartphones and tablets).
And also Google handles around 92 percent of all internet searches, and YouTube is used by 73 percent of the US population. In addition to these various sources, it records every click, tap, query, and movement made by each user.
In order to meet all this demand, Google has many different Data Centers across the Globe capable of processing and delivering a user’s search with the results in less than a second.
The AI maturity level
Having that in mind, it is necessary to have the Framework for AI adoption to help technology leaders who wish to establish an adequate AI capacity to utilize the power of AI to effortlessly and intelligently improve and optimize their business.
According to the organization’s level of maturity in AI adopting, in our view, there will be greater or lesser proficiency with the AI use cases, which will, in turn, have a direct impact on the Data Monetization process.
Thus, the higher the organization’s AI maturity level, the more AI use cases there will be, and the greater the re-use of data for different purposes, and as a result, the greater the diversity of data required for the AI use cases.
Building the basis of an AI Strategy
The adoption of an AI Strategy should be approached from two perspectives: bottom-up and top-down. First, the bottom-up approach is used to understand what people on the front lines think can be improved. Then, the organization’s board of directors operates the top-down approach to filter and gain a clearer vision and frame the final version.
It is possible to create AI use cases from a continuous process of understanding how business problems that require specialized knowledge within a company can have their outcomes improved by using AI models to generate value for the organization, a process known as machine learning.
An early assessment of the organization’s AI Maturity level is recommended to identify areas for improvement and focus on developing a roadmap with specific milestones to be met.
We believe that this should be done precisely to raise the organization’s excellence in deploying artificial intelligence use cases to achieve improved performance, measurable efficiency gains, and enhanced business outcomes.
References:
- Data Governance
- Data Monetization:
- Google Cloud’s AI Adoption Framework
- Model OPS and MLOps
