Whether it is to calculate credit scores, detect fraud or estimate insurance premiums, artificial intelligence has already entered the financial sector in Quebec. The Autorité des marchés financiers (AMF) must now “establish a model framework” for it to be used in a responsible and ethical manner, concludes a report made public on Monday.
“We must obviously take advantage of the opportunities offered by technological innovations,” said AMF CEO Louis Morisset, who announced the publication of this report commissioned and funded by his organization during a virtual conference. “But we must also and above all remain vigilant with regard to the negative consequences that are still little known, or poorly understood, which could materialize in the markets or with certain groups of consumers. ”
Urgent need
Led by Marc-Antoine Dilhac, assistant professor in ethics and political philosophy at the University of Montreal and instigator of the Declaration of Montreal Responsible IA in 2018, the committee submitted 10 recommendations, 3 of which are addressed directly to the AMF.
“The need to supervise the use of AI in finance is becoming urgent,” we write in the preamble. Regulators must exercise their authority to ensure that the use of AI does not skew financial markets, threaten fair competition, and harm consumers. ”
Consumers, we note, “feel particularly vulnerable in a digital and algorithmic environment that they do not always understand and do not master”.
As a first step, indicates the report, the AMF must establish the basic principles of this responsible IA which will then be able to inspire the codes of ethics of the financial institutions, by defining in particular “the unacceptable practices”. We must also, after consultation with the financial sector, define the sources and types of data that it is legitimate to use. For the third recommendation, we suggest that the AMF train the players and develop education programs for consumers.
The other seven recommendations are addressed primarily to financial institutions, requiring them to ensure, for example, that their algorithms do not create discrimination, that their systems are robust and that they can be audited.
It is emphasized that data collection “respects the autonomy of consumers by communicating all the information required to consent in a free and informed manner. […] in clear language, and respecting the diversity of lifestyles ”. Basically, this is what Europe has been trying to establish since 2016 with the General Data Protection Regulation. The CEO of the AMF noted, during his speech, that social networks and connected objects “produce masses of data about our choices and our journeys”.
Encourage reflection
According to Louis Morisset, this report by the committee headed by Professor Dilhac firstly wants to stimulate “dialogue between interested stakeholders on the responsible deployment of artificial intelligence in the financial sector”.
The CEO hopes that it will generate reflection within organizations, ensuring that the AMF has started its own in 2015 on this subject.
“We firmly believe that the use of artificial intelligence in finance will only bring lasting benefits if it takes place while respecting the values of Quebec consumers,” he said during his speech.