Like a highway without speed limits or police officers to control traffic, generative artificial intelligence (AI) is creeping into classrooms and university work, raising several ethical and educational questions detailed in a voluminous report published Thursday by two organizations governments. The Superior Council of Education (CSE) and the Commission on Ethics in Science and Technology (CEST) are therefore urging Quebec to guide the use of this technology in the province’s CEGEPs and universities.
“If we had all the current automobiles and means of transportation, but there were no markers, there were no regulations to manage all that, can you imagine what that might look like? » launched Tuesday at Duty the president of the CSE, Monique Brodeur. This would be inconceivable, she notes. However, generative AI represents risks just as significant for the school and university environment as a deregulated highway, even if this technology, which has experienced rapid development in recent years, remains “abstract” for a good part of the world. population.
“The problems are just as big as in relation to motor vehicles, if they are not even more serious,” continues Mme Brodeur, according to whom generative AI can pose “several threats”. Thus, “rapid supervision” of Quebec, through legislative and regulatory reforms, is necessary, she continues.
Caution
The risks posed by generative AI – but also the possibilities it offers – the CSE and the CEST have documented at length in a 138-page report entitled “Generative artificial intelligence in higher education: educational and ethical issues “, that The duty was able to consult under embargo. This document is the result of a consultation carried out last year, which resulted in the collection of 26 submissions, notably from colleges, universities and federations representing teachers and students.
The slogan that emerges from this document, which includes around twenty recommendations intended for Quebec and higher education establishments? Prudence.
“We consider that what is preferable at the moment is really a cautious approach, so that we can see what the short-term and medium-term needs are for the different stakeholders and so that we can be especially able to document them correctly,” underlines the president of the CEST, Luc Bégin.
Indeed, there are already multiple tasks that generative AI can carry out in a university environment, both for the benefit of students and professors, who are increasingly using it. Conversational robots can, among other things, develop lesson plans, correct exam papers, generate computer code and produce short essays.
However, the use of AI for these purposes raises a series of concerns about the risks of plagiarism, fraud, but also misinformation posed by this new technology. The latter could also contribute to spreading information conveying discriminatory biases, particularly against visible minorities, notes the document.
The latter also notes the risks that conversational robots cause a “cognitive discharge compromising the acquisition of knowledge or skills deemed necessary to achieve learning objectives” by several students.
“We can think of memorizing information, but also of skills such as the ability to conduct in-depth analyses,” underlines Mr. Bégin. From the moment the student uses generative AI, will she, and in a free and very open manner, continue to develop this type of capacity? »
Fund research
Thus, the authors of the document recommend that teachers make their students aware of the importance of not placing blind trust in generative AI and thereby diversifying their sources of information. For its part, the government should fund university research on the various risks posed by this new technology and the measures to take to adapt our teaching environment to it. These research efforts would be overseen by a “collaborative national consultation body”, which would support the government in developing regulatory guidelines to govern generative AI, the document indicates.
“We are going to need studies which do not just describe the phenomena, but also which will provide us with relevant information to act responsibly and diligently because this issue is really evolving at lightning speed,” notes Mme Brodeur, who calls on Quebec to show “leadership” in this matter.
Specialized training on AI as well as its uses in the context of learning and teaching funded by Quebec should also be offered to both students and teachers so that they have the necessary knowledge to use this judiciously. technology, recommend the authors of the report.