Predictive Inequality in Artificial Intelligence Modernism
Currently, in the world there are 12.5 billion internet connected devices such as smartwatches, security surveillance systems and smart home appliances, this has already surpassed the number of human beings on the planet that is 7.53 billion people. This market is rapidly growing as at least 26 billion of internet connected devices are projected to be in use globally by 2020. These devices are used by consumers in their daily life, thus, the term Internet of Things (IoT).
The beauty of IoT is that even small device like a smartwatch is actually connected to much larger systems and networks. They are complex systems and vulnerability at any point can cause damage at both individual and institutional level. Smart products are inherently designed to share information and communicate with other connected devices. This data may seem innocuous on its own but when put together and analysed reveals a rather accurate profile for the user.Therefore, it is imperative to make these everyday products smart and trusted.
These recent technological innovations such as Artificial Intelligence (AI) and blockchain are giving us a chance to remake the world. AI is a very advanced technology affecting both developed and developing economies. Hence, the important question is how we can ensure that AI is used to create a better world that uplifts everyone. AI has the potential to create a trampoline effect especially for the developing economies that helps them bounce upwards skipping some steps and using automation to elevate their overall socio-economic status. The economics of this digitization is compelling.
However, this fourth industrial revolution brings its own set of challenges. Some of the major concerns are not just regarding security or adaptation of new technology but also a social impact that comes about as a consequence of these high-tech advances. It is essential to understand the benefits and risks of technical cultural changes. As technology not only supports operations but acts as a powerful tool to drive innovation and influence future strategic decisions in almost every sector.
An unfortunate consequence of this digitization highlighted at the World Economic Forum late last year, was increasing inequalities. One of the biggest reasons for connected devices being biased is because of the data they use. This digitization primarily relies on tonnes of data. The more historical data that is fed to Artificial Intelligence, Machine Learning and Deep Learning, the better the future models will work. Majority of people are unaware of the value of the data they produce and its role in Machine Learning market. Thus,we need to understand that Machine Learning only processes the data input and if the data is biased so will be the Artificial Intelligence system.
Current research has highlighted racial and gender bias in AI. It is kind of a chicken and egg situation. Better data makes a better product/service or better product/service generates better data. For example, in addition to worrying how safe self-driving cars are, an academic study found that self-driving cars were more likely to hit dark skinned people. This is the case because automated vehicles, based on the data provided, were better at detecting pedestrians with lighter skin tones. This provides evidence of how the human biases seep into decision-making algorithms.
Moreover, machine learning is also putting women at a disadvantage in many cases. For example, the use of AI to detect heart attacks based on general symptoms misdiagnoses heart attacks in women as heart attack symptoms differ in men and women. And machines have been fed data related to mostly men, thus, it fails to diagnose women accurately creating an unintentional yet a serious problem.
There are many solutions to tackle these challenges but one common answer is awareness for consumers, developers and policymakers.
The first step is to explain complex products to consumers. They may understand how the device connects to internet and functions but they often do not understand what is happening behind the scenes. They should be explained basic practices for example, how data is collected, stored and processed and how vulnerable the system is to biases in a concise and clear manner. Moreover, in addition to the main producer, there are other players in supply chain for IoT services. This information needs to be evident to clear any ambiguity as to who is accountable for which part of the supply chain.
Another reason for this bias is over representation of men in designing these technologies. It was reported that only 22% of AI professionals globally are female. The digital transformation has the potential to unlock solutions to global issues and inclusion of women and minorities will help create better systems and build trust. Organizations need to diversify their teams,women need to be builders and end users of the AI enabled products and services. Private sector needs to attract women, minorities and people from diverse disciplines for jobs in AI. They can also invest to re-skill existing human resources in the organization and promote use of AI devices at workplace.
Lastly, governments also have a role to play. Government is like a conductor who instructs an Orchestra. It is his duty to make sure that the musicians are playing in synch but without forgetting that the melody is played for the audience, the citizens in our case. They need to formulate and implement stringent regulations that protect all citizens but at the same time does not hamper innovation either. One of the reasons for limited data is that only half of the world has access to internet with developed countries being far ahead of developing. If we are not careful, institutionalized bias has a risk of potentially more unequal world as only the people with technical abilities and access to basic digital literacy will thrive. With every new website and smart innovative feature, the divide between those who are online and those who are not broadens. This proves the famous notion that the future is here but not everyone has access to it.
In this moment of technological change, leadership matters, at both the government level and in private sector to help the nation not only be a recipient of AI era but to embrace and adopt it for better future for all of its citizens. There is still plenty of room for developing economies like Pakistan to learn and develop AI systems for the upcoming AI powered world. Thus, it is imperative for policymakers and private sector to work together to ensure that the other half is not left behind.
This article was originally published at:
The opinions expressed in this article are the author's own and do not necessarily reflect the viewpoint or stance of SDPI.