A research laboratory, LaborIA, was launched on Friday, under the aegis of the Ministry of Labor, to better understand the effects, still poorly understood, of artificial intelligence (AI) on working conditions, recruitment, skills. ...
To discover
SIMULATOR - Are you getting paid well?
Everything you need to know about skills assessments
This 5-year program, the result of an agreement between the ministry and the National Institute for Research in Digital Sciences and Technologies (Inria), will be operated by the Matrice innovation institute, a start-up incubator -up and trainer of job seekers in web development and data processing.
Read alsoShortage of manpower in the field of artificial intelligence
"
The goal is to concretely test AIs in a work context: transformation of careers and recruitment, tools for social dialogue, automation of tasks and job development, impact on low levels of qualification
", explained its director François-Xavier. Small during its presentation.
It will take place in three phases: construction of a barometer on AI at work based on a survey carried out among companies between November 2021 and April 2022;
formalization with companies of concrete projects;
launch at the end of 2022 of six-month experiments which will make it possible to make recommendations that can be extended to other companies
"Find meaning"
During a round table, the various speakers insisted on the importance for the user to understand "
the values on which the algorithm was designed
", which can "
impose management methods
" or "
muddy the waters between l 'human and non-human
'.
Faced with the explosion in the amount of data, AI is already used in many economic sectors or by certain public actors such as at Pôle emploi.
"
But end users must find meaning in it
", for example, "
doctors need to understand how the data will improve therapeutic management, otherwise it has no value
", underlined Salima Benhamou, who follows these questions to France Stratégie.
Read alsoWhy artificial intelligence is far from replacing the radiologist
Sociologist at Icam, Yann Ferguson warned against the reproduction by the algorithm "
of biases linked to the community which built the code
" or "
linked to the data set transmitted by the company
".
And "
the risk of loss of expertise or critical thinking if we place extreme trust in him
."
“AI asks lots of questions in the workplace, it's dizzying.
But it is better to research the issues upstream than to discover their effects a posteriori ”, concluded the Minister of Labor Elisabeth Borne.