Already 3 years ago, in June 2016, The Drum, an American magazine specializing in marketing and digital, trusted Watson. This artificial intelligence (AI) platform designed by IBM was to partially replace a team of journalists and write in-depth reports in a historic issue symbolizing the intersection between the intellectual capacities of men and the efficiency of intelligent computer systems …

In view of this experience, undoubtedly a source of future editorial innovations, fundamental questions arise: what place will humans occupy in the editorial offices of tomorrow? Will content creators be machines? What added value can Tech bring to a sector that lives more on letters than numbers? How far are men ready to let disruptive technologies, and more particularly AI, interfere in their professional life?

Training & Co'm

Where are we really in AI?

In 2019, AI is considered to be the science whose goal is to make a machine do tasks that humans do using their intelligence. In practice, this covers a wide variety of missions such as identifying images deemed “violent” in a video, detecting comments with racist overtones on a Facebook page and instantly deleting them, translating an entire article into Spanish into French or even , as the Associated Press agency already does, write financial reports and publish them directly on its web page.

Marc Damez-Fontaine, Director of Data Intelligence at PwC, comments: “Many people wrongly associate robotization with AI. These are two completely different concepts. Robotization is the automation of a process: a machine performs tasks that Man orders it to perform, and it improves thanks to Man who integrates new programs into the machine. Its level of autonomy is zero.

AI, on the other hand, corresponds to an autonomous learning process: the machine performs tasks by imitating the tasks performed by humans beforehand. She develops her model on her own, learning from her mistakes and can adapt to the environment around her. “He adds:” to develop, it feeds on data that represents human tasks. The more data a machine records, the more it is able to adapt to its environment and make decisions similar to what a human being could make. ”

To date, there are two types of AI: weak AI and strong AI.

Weak AI essentially aims to capture as faithfully as possible, using algorithms, the result of specific behavior expected, without any form of improvisation. This technology only performs scheduled tasks. It does not evolve. In other words, she does what she is asked to do, nothing more.

Examples: propose a targeted advertising campaign by applying preprogrammed strategies on its server, write dispatches based on an in-depth semantic analysis, or even respond orally to questions asked by men thanks to the quantity of questions and answers recorded in the machine.

Weak AI is very efficient in its field but remains confined to it because it is content to apply instructions dictated by Man. Today, this is the form of AI that researchers are mainly developing.

To the
difference from weak AI, strong AI corresponds to a computer program
able to reason, learn and even solve problems … As if
she was “intelligent” in a way. These machines are capable
to learn alone from lived situations, just like a child
who ends up understanding for himself the logic of a game by trying to
to deceive, to correct the course then to succeed. In 2019, strong AI is located
still in the experimental stage.

Weak AI at the service of content customization

The fields of application of AI in the communication sector are vast: automation of business processes and customer relations, monitoring and listening to social networks, verification of the authenticity of the information retained, predictive analysis of the success of editorial strategies , video development and post-production, voice and conversation assistants, automated writing, personalization, recommendation and optimization of content distribution, “tracking” of upcoming trends, etc.

Now, any action is planned, calculated, deployed and measured before being recycled, readjusted and reprogrammed. Thanks to – or rightly say, some purists – new technologies, the share of human improvisation is reduced. The dispersal of teams is banned. Algorithms advise, Men think and machines execute without wasting time.

The era of augmented communicators who rely on Tech to be more efficient, precise and quick in their work is by no means science fiction. This is the present. And for some, it is already a thing of the past.

Supporting examples: since 2015, journalists from the famous French daily newspaper Le Monde have hired the services of a Syllabs robot-editor to write informative articles during periods of departmental and regional elections.

Thanks to the Heliograf platform, developed in 2016 for the Olympic Games, The Washington Post uses AI to cover second-rate events, such as local student sports, whose audience is too small to mobilize a human journalist. Since that same year, the Finnish television channel YLE has trusted the Voitto bot to publish more than 100 articles and 250 images on its website each week.

Finally, more recently, in 2017, Benoît Raphaël, former journalist who became an expert in digital innovation, developed a project called Flint and Jeff, two robots which, thanks to the personal data provided by the users, offer them a daily newsletter with quality personalized articles, handpicked and which correspond to the evolution of their tastes and their different areas of interest.

These few examples are far from isolated cases. With viral communication campaigns, spectacular fundraising and praise from opinion leaders on the progress of this disruptive technology, AI is triumphing, more and more companies are integrating it into their business model , and, logically, research and development in this area is intensifying from year to year. At this rate, experts say, by 2030, the AI ​​market is expected to brew more than $ 15.7 trillion worldwide.

Data, the new black gold for businesses

To date, AI research has focused primarily on three areas: automatic environment perception, combinatorial problem solving, and machine learning – more commonly known as machine learning.

David Cortés, Director of Data Intelligence at PwC, comments: “Today, what convinces scientists is only machine learning which is a small part of what will make up the AI ​​of tomorrow. Its operation, in a very simplified way, is based on statistical approaches to give computers the ability to learn from data.

In this context, the more machines feed on data, the more likely they are to be efficient. Hod Lipson, professor of mechanical engineering at Columbia University, adds, “Data is oil and algorithm is the engine. We used to program computers. Now the machines learn themselves. We just need to feed them with data. In other words, without data, doing AI is impossible!

In one
increasingly tech-savvy world, scientists question
: we are moving towards a “human-driven”, “data-driven” society
or a little of both? Who masters what? What controls who? who
understand who?

According to an international study by The Drum and Sysomos, only 46% of communication and marketing professionals believe they have a good understanding of the challenges of AI and 37% are already carrying out pilot projects in their field to better understand the exact potential of this disruptive technology developed over 70 years ago. Technological progress awaits no one and yet the adoption of AI by the vast majority of communicators is slow, very slow, too slow.

The risk here is significant: seeing a digital divide intensify between those who embrace Tech and use it to revolutionize the marketing and communication sector, and those who, as best they can, follow progress from afar and try not to be outdistanced by competition too quickly.

Modernization of necessary tools and methods

The communicators of tomorrow will have to be even more connected, master new digital tools, and know the codes and habits of the target audiences. They will have to be technically able to rely on intelligent platforms to make better decisions and automate recurring tasks, personalize all the approaches carried out with “stakeholders” to offer them a unique user experience. The communicators of tomorrow will have to develop a certain emotional intelligence in order to humanize their actions as much as possible, to be present in circles of influence online and offline, to bring together and gather the crowds. They will have to create emotions and, in an era where all information drives out another with a snap of their fingers, have a definite impact without occupying the center stage 365 days a year. To be influential in the era of “I know everything and I want to know everything” is also to understand that speech must be rare, precious and mastered.

Despite the
makes the societies in which we live increasingly virtual
and artificial, the real, the authenticity, the proximity, the human and the
emotions it releases remains the first success factor of a campaign
communication. Man is a means of attracting attention. Man is a
goal. Man is a lever
growth. Tech is just a tool that allows him to achieve his
purposes, nothing more, nothing less.