Artificial Intelligence allows machines to learn, both supervised and autonomously. The proliferation of Cloud technologies, the digitalisation of images, texts and audios and the development of IoT (Internet of Things) have made it possible to gather the large volumes of information that machines need to learn.
In this way, machines, through Artificial Intelligence, acquire the ability to find patterns and relate data, events or variations even imperceptible to the human eye; calculate what is going to happen in a certain area and even provide answers to questions thanks to data analytics, with the potential that this has to help us find solutions to some of the problems of our society.
Artificial Intelligence is already present in our lives much more than we realise.
Some everyday of Artificial Intelligence examples are the algorithms that recommend content on video-on-demand platforms, those that prevent and detect fraud by identifying anomalous use of bank cards, or those that recalculate the route in the car’s GPS based on traffic conditions.
Artificial Intelligence has also demonstrated its capacity in areas such as industry, where it prevents failures and breakdowns in machines and systems to avoid incidents or unforeseen stoppages; health, where it has numerous applications in both the diagnosis and treatment of diseases such as Alzheimer’s or cancer; or education, where it can anticipate school dropouts, detect talent, or personalise study plans based on the abilities and individual needs of each student.
Towards a responsible Artificial Intelligence
The benefits of Artificial Intelligence are therefore enormous. It allows us to reach far beyond what our human analytical capacity makes possible and opens up great opportunities in the use of data by companies and organisations.
The growing importance and influence of data in our lives makes it necessary to develop responsible Artificial Intelligence in which algorithms pivot around three essential principles: ethics, transparency and explainability.
- Ethics: as algorithms acquire the ability to make or influence decisions, they need to respect social norms so that they are fair, inclusive, diverse and respectful of privacy.
- Transparency: to avoid algorithms being “black boxes” in which we do not know what happens, we need to know how they are applied and how they work, being able to access the data sources used and the mathematical formulae employed.
- Explainability: we need to be able to understand the “behaviour” of the algorithm, what results it is generating and why it is generating them, or why it makes a decision or arrives at a particular deduction and not another.
Ensuring that the data that will be used to train and teach the algorithm are free of bias and are shaped in a fair manner, aligned with human rights and in line with the rule of law, especially when dealing with personal data, is critical for an algorithm to be ethical.
Principles of ethics and transparency
In this sense, the European Union’s regulatory model is oriented in this direction and public bodies and large companies are focusing their efforts in this direction. One example is Telefónica, which published its Ethical Principles on the Use of Artificial Intelligence in 2018.
Children need to be taught about computational thinking, algorithms and Artificial Intelligence.
The key to complying with the aforementioned principles is to improve the population’s level of knowledge about Artificial Intelligence by investing in education in this area. Children need to be taught about computational thinking, algorithms and Artificial Intelligence, just as they are increasingly trained in programming and computer science.
These principles of ethics and transparency are critical to building a responsible and inclusive Artificial Intelligence that fosters equal opportunities and drives economic and social progress. In short, Artificial Intelligence at the service of people, which contributes to building a better society.