News
Robots with Artificial Intelligence are no longer characters of science fiction. They are deeply integrated into the routine of millions of people and thousands of companies and industries, from the most diverse fields and sectors. Who hasn't talked to a chatbot or been assisted by a robot on the phone? In times where communication and interaction are drastically being altered and redefined, with robots with artificial intelligence assuming increasingly human faces, personifications, and expressions, very human questions are brought to light. Is there any ethics that controls the actions and “thoughts” of these intelligent systems? To what extent does the human factor apply to them? Who should be punished in case of misconduct? These questions are based on biased decision-making and some controversies caused by robots or AI systems, such as reproductions of biased speech, for example. For the people who encountered these situations, the criteria used by these intelligent systems did not seem clear enough. Some companies responsible for the systems tried to bypass the issue, but until then there was no centralization of the theme, internationally speaking. In 2019, the European Union took the first step to create a globally applicable model that would encompass all companies developing artificial intelligence systems. ALSO READ: • Computer Vision: a complete guide! Check it out here!
• Artificial intelligence in the electrical sector
Speak with one of our specialists and discover how Pix Force can transform your business
International model regarding artificial intelligence robots
With this, the EU established seven points to define whether an artificial intelligence is trustworthy or not: • Human oversight: Every Artificial Intelligence must allow for human intervention in its operation. Instead of diminishing human autonomy, AI must ensure the freedom and equality of human beings through its actions;
• Robustness and safety: Systems must be reliable, secure, and have the firmness and solidity to handle adversities and errors or inconsistencies during their operation. This is one of the biggest challenges for developers today: creating algorithms capable of better understanding complex issues and responding quickly to unforeseen and non-standard situations.
• Privacy and data governance: Artificial Intelligence must be transparent and provide full access to users' individual data, as well as guarantee their privacy and security in situations where personal information could be used to harm or discriminate against them.
• Transparency: All actions and decisions of the system must be clear, readable, and understandable to the public. Even if its programming is complex and technical, its activities must be translated so that users are able to fully understand its essence.
• Diversity, non-discrimination and fairness: Systems need to encompass and understand human beings. This includes understanding and accepting diversity in all its aspects. In this way, AI must value inclusion, acceptance, and accessibility for the most diverse audiences and human beings.
• Social and environmental well-being: Artificial Intelligence must be a resource and a tool to assist human beings and humanity. It must be employed to produce significant and beneficial social changes for the population, without neglecting social, ecological, and sustainable responsibilities.
• Accountability: In case of errors or malfunction of the systems, there must be a way to account to the users. In this way, competent developers can be held responsible for the results and consequences of the AI's malfunction.
Application and future of the model
The model was created with the help of 52 specialists in the field, including academics, sector organizations, and companies such as Google, Bayer, Santander, and SAP. As this is a model still under testing, the points raised do not have the force of law. However, they should serve as guidance for developers of Artificial Intelligence platforms and systems, despite there being no punishment or restriction in case of non-adherence to the model. The European Union's objective is, however, to expand and improve the model beyond the countries that make up the bloc. The European Commission intends to take the agenda to Asian countries like Japan and Singapore, major exponents in this technological branch and important suppliers and exporters of intelligent systems globally. In addition, the agenda should also be discussed with the G7 and G20, of which Brazil is a member. The pilot project will be evaluated in 2020 by Artificial Intelligence specialists so that the next steps can be discussed. Each of the points will be analyzed, separately and together, for a more accurate perception of the viability and applicability of the model and the identification of improvements and changes.
Why is this important?
The human routine has already incorporated intelligent systems so deeply and naturally that it is almost no longer noticeable when we are dealing with other humans or not. In this way, procedures need to be reviewed to adapt the activity of these systems to the accepted standard in human society. Just as humans must follow rules of behavior and conduct, so must robots. Since all intelligent systems are created and developed, from the first line of code, by human intelligence, there needs to be a standardization in the line of conduct to be followed by these programmers. Thus, errors are reduced and scopes are expanded. Cases like Microsoft's artificial intelligence that reproduced biased and sexist comments on Twitter should serve as an example. This way, these errors are not repeated and can be taken into consideration in the future. Only then will robots with artificial intelligence stop being viewed with suspicion, ensuring safety and comfort for humanity, as is their purpose.
Fabio Caraça
Fábio Caraça is the Chief Growth Officer at Pix Force. He leads Pix Force's transformation into a scalable SaaS operation, combining strategic vision, culture, and high-impact execution.
