AI and data, towards more responsibilities

For a few decades, data and Artificial Intelligence have become omnipresent in our daily lives, but how can we move towards an ethical and responsible use of this data?

Context and issues

Before addressing the concrete issues of responsible data, we suggest to quickly go back to the context that led us to ask this question.

The advent of digital capitalism in a few dates. First of all, the beginning of the 21st century saw the emergence of a first phenomenon: the birth of a new kind of companies: startups. Most of these new models offered digital services that were considered innovative and promising at the time. This phenomenon attracted a plethora of investors, all looking for new business opportunities.

However, in 2001-2002, the unpredictable crash occurred and the Internet bubble burst. Several questions emerged then: why did this new model explode so quickly? And above all, how to explain such a sudden collapse? The immaturity of a market with no viable business model was the main reason given.

GAFAM, the foundations of a sustainable system

Despite its tumultuous beginnings, digital capitalism has not said its last word. Indeed, the rise of GAFAM (Google, Apple, Facebook, Amazon, Microsoft) marks the beginning of a second phase. This is illustrated by the digital exploitation of the human experience, initiated by Larry Page and the Google model, among others.

The advent of Big Data

Even before Google had the technology to achieve such a result, the vision was there. The democratization of increasingly sophisticated digital devices and the exponential development of social networks have transformed the place of data in modern societies forever.

Our privacy is now used as raw material, our digital actions are collected, recorded, monitored and finally transformed into behavioral data. All the players in the sector have copied this model, starting with the most important ones, Facebook and Google. However, the responsible use of data and AI does not seem to be at the heart of these players' concerns.

AI, a tool for data transformation

Today's digital factories no longer have large chimneys... They rely on the collection, refinement and exploitation of data. The arrival of artificial intelligence in the operational world marks a real revolution in the use of data. Based on mathematical concepts that are sometimes empirical, AI feeds this data to create "products" that support an economy based on the analysis and prediction of human behavior.

These new markets for online advertisers and others are at the origin of more than 90% of Google and Facebook's revenues. These companies have succeeded in transforming personal information into the black gold of the 21st century, and thereby, are continuing to expand their commercial empires.

What is the place of ethics in the New World?

These observations inevitably raise serious reflections about ethics. Indeed, the control of data and AI will condition the daily life of future generations.

A notion that struggles to be heard

However, at this stage, no one seems to be mature on the notion of ethics and responsibility of data exploitation. The major international powers seem to be deeply convinced that the sector must be regulated, with the European Union in the lead. Regulations exist, and hundreds of draft bills have been introduced in recent years. The legislator's goal is to protect online privacy and digital disinformation.

The use of AI is nowadays torn between three factors: its power, its overselling and its denunciation. Nevertheless, it remains a formidable tool as long as it is equipped with essential principles of responsibility:

  • Principles of ethical responsibility by protecting itself from the biases it is likely to cause or accentuate. The latter could lead to major upheavals, injustices and inequalities.

  • The need to ensure the transparency and explicability of the functioning of algorithms and the data from which they have been trained.

  • Principles of accountability around the data that AI algorithms use.

  • Principles of accountability against the abuse of business ethics with false promises or fraudulent and biased exploitation of AI concepts and data.

Data and AI management: how to act as a responsible company?

From a company's point of view, what does the notion of responsibility around AI and data mean? How can we adopt a responsible behavior? The answer lies in the question. At a minimum, it is about finding a responsible and ethical balance between the interests of the company, the interests of its ecosystem, and the legal frameworks that allow to frame behaviors and rules.

The company using these concepts must clearly express its responsibility for the collection and use of data, as well as for the AI principles it may be trading:

The responsibility lies primarily in the use and thus the collection of data. Rather than pursuing a strategy based on the profusion or collection of data at all costs, it is more important to focus on the quality (origin, freshness) and targeting of the data by thinking upstream about the uses.

The responsibility extends to giving more importance to the ecological footprint of data storage and processing. The mad rush to collect and store data multiplies exponentially the amount of data stored in data centers.

It continues with collective education: educating people on usage, ethics, societal and legal responsibilities, training staff in charge of data collection so that they have the technical and legal knowledge to ensure their integrity (ethics).

Classification and protection of data, defining its criticality and scope in terms of impact, and implementing more or less extensive security measures are all guiding principles for which the company also has a responsibility.

Responsible data and AI, towards a new era

Having a responsible data management policy also means being transparent in the monetization of data. It means agreeing to make the explanation of data use explicit and accessible. In the end, it is above all a matter of being part of an ethical approach via a logic of reciprocity. Each contributor in the chain should contribute to the value created and share the benefits.

To conclude, we are perhaps at the dawn of a paradigm shift. Indeed, today's world seems to be moving towards a more united exchange and sharing of data. Data and its exploitation probably have an important role to play in building tomorrow's world in a more sustainable way, without calling into question the economic model of companies in the development of their activity.