By Stéphane Amarsy,
Chairman of the board of Splio + D-AIM

The increasingly widespread use of artificial intelligence in our everyday lives raises questions about its impact on both individuals and society as a whole. Given the potential for current and future excesses, many countries have been trying to establish regulatory frameworks.

At the same time, people across the globe have contributed towards establishing ethics committees to guide future developments of AI and its components. Yet, the question of how to bring diverse interests together remains complex. Advancing digital regulations and setting up ethical monitoring is a Sisyphean task for the modern era.

 

 

Automatic decision systems: between hope and fear

Automatic decision-making systems, which use statistical learning techniques to process data, raise some legitimate hopes across various sectors such as sales, law, administration, finance, industry and medicine. But when it comes to preventing these technologies from being misused or abused, the key actors involved shouldn’t bear sole responsibility. Potential risks include discrimination, arbitrary decisions, liability, the excesses of technological development for its own sake and the biases associated with data interpretation that ultimately impact the algorithmic results.

 

 

The use of algorithms must be transparent and ethical

Algorithms need a legal framework

 

How can people be protected from unfair decisions that affect them directly? For example, how is your credit score calculated? Does it come only from you or is it also influenced by what’s happening around you geographically? Widespread use of algorithms without transparency or ethics will lead to extreme societal fragmentation. That’s why they must be legally regulated and developed ethically, which is not always the case. Data scientists have a responsibility to ensure they are fair. And when it comes to the data used, along with metrics and results, everything must be audited and clearly explained. This will inevitably lead to better algorithms.

 

In 2017, the CNIL (France’s National Commission on Informatics and Liberty) initiated a discussion on the topic of “Digital Ethics: Debating Algorithms,” which led to the publication of a report. At the same time, a commission led by Cédric Villani published a report whose objective was to “Make Sense of Artificial Intelligence.” Both reports focused on the ethical questions raised by everyday uses of AI in all aspects of life. Many countries around the world have already begun work on this issue and launched initiatives. It’s important to question the impact that the widespread use of AI has had on decisions that directly influence people’s lives, whether in banking, insurance, health, employment, the legal system or law enforcement. To this day, a great deal of work remains to be done.

 

Addressing these issues requires specific technical skills – to understand how algorithms work – along with a critical examination of discourses and , as well as legal, societal, political and philosophical expertise. While this is undoubtedly an enormous challenge, it can be achieved if we provide ourselves with the right tools.