Innovative uses of artificial intelligence are increasingly visible on a daily basis and its benefits are increasingly tangible. AI is now entwined in all the processes of economic life: optimization of industrial processes, human resources management, marketing and customer relations, financial analysis, medicine and public services, to name but a few, and is no longer the preserve of disruptive innovations in cutting-edge industries.
However, while ever more players (consumers, public authorities, managers, employees, data experts, etc.) are realizing the benefits of artificial intelligence, particularly as a lever for economic and technological innovation, those who fear its possible abuses are many. We therefore need to question the ethics of artificial intelligence for the years to come, and the challenge is to define the shape of Responsible AI.
According to the European Commission, "artificial intelligence (AI) systems are software (and possibly also hardware) systems designed by humans that, given a complex goal, act in the physical or digital dimension by perceiving their environment through data acquisition, interpreting the collected structured or unstructured data, reasoning on the knowledge, or processing the information, derived from this data and deciding the best action(s) to take to achieve the given goal."
To put it more simply, AI represents any technique allowing machines to imitate human intelligence.
AI was studied in 1950 by Alan Turing, a British mathematician known for his test based on the ability of a machine to hold a human conversation. It was then formalized in the United States in 1956 by Marvin Minsky and three other researchers. It therefore constitutes a set of sciences, theories and techniques that will enable the interpretation of complex data collected on a large scale.
Thanks to machine learning, deep learning, big data and artificial intelligence, many AI applications can facilitate people's everyday lives. Artificial intelligence now occupies an increasingly important role in our activities and is a lever for innovation and progress in many sectors.
Although advances in artificial intelligence are making a major contribution to improving working methods and making life easier for users, how it is perceived is mixed and it remains a subject of uncertainty among the general public. Thus, only 31% of French people believe that the development of artificial intelligence systems has more advantages than disadvantages (IPSOS, 2022).
Where employees are concerned, only 40% expect a positive impact from AI, while 43% fear job cuts following the adoption of AI in their company (BCG GAMMA & Malakoff Médéric Humanis study, 2019). Employees and, in particular, data experts who create or use AI systems are looking for meaning in their career paths, and are expressing their desire to get involved in companies that share their ethical concerns.
Left unregulated, artificial intelligence can also give rise to ethical risks:
The major challenge of AI is that it can reproduce and amplify already existing social biases. In practice, an AI system may favor a certain class of people over others. Conversely, a monitored AI makes it possible, for example in the context of a recruitment process, to erase the discriminating information, such as ethnicity, gender and age of candidates when processing résumés.
When an algorithm generates a result or a decision that is difficult to explain, this is referred to as the "black box" effect. If it is not regulated, it can therefore have an impact on end user protection and even the respect for fundamental rights. The implementation of ethical, transparent and responsible AI is therefore essential.
Artificial intelligence systems must be designed to support and complement the skills of human beings, without interfering with their decisions. To make sure that AI does not coerce or subordinate humans, it makes sense to ensure that AI systems and processes remain under human supervision and control.
Although the idea of a standard comes up regularly in France and Europe, there are currently very few operational solutions for companies wishing to have a responsible approach to the deployment of their algorithms. On the other hand, consumers and employees also have few means of judging the ethics of the companies for which they work or with which they interact on a daily basis.
The European Union has initiated a major debate on the framework for artificial intelligence systems through several legislative proposals (AI ACT). Positive AI aims to support this debate in two ways:
By anticipating regulatory requirements on behalf of companies, by providing tools and a label;
By participating in the debate through the strength of its proposals
Given the urgent need and the high stakes involved, it is vital not to wait until a European regulation comes into force, but to act now to offer solutions to organizations wishing to improve and move toward Responsible AI.