91�鶹������Ʒ����

The survey was prepared for the Ministerial Roundtable on Generative AI and Education, organized by UNESCO on 25 May 2023.

The result illustrates the uncertainties in responding to the sudden emergence of newly powerful generative AI applications that can produce human-like output, including summaries, essays, letters, computer programs, art, and more. The technology is also currently capable of scoring top marks on major standardized tests, including university entrance examinations and assessments to credential professionals, including doctors and lawyers.  

The vacuum of guidance shows that education systems are working to catch up with the new technology.

Policies governing the educational use and misuse of novel digital technologies often take shape at the institutional level before being inscribed at higher district, state, and/or national levels. The few institutional-level policies surfaced by the UNESCO survey indicates that education systems are still finding their balance and deliberating their responses. More system-wide policies – applying to large numbers of schools and universities in national and sub-national contexts – will likely take considerably more time to formulate and coalesce.

“The survey results show that we are still very much in the wilderness when it comes to newly powerful generative AI and education,” said Sobhi Tawil, the UNESCO Director for the Future of Learning and Innovation. “Institutions are not yet providing guidance or direction.”

While schools and universities appear to be taking their time to make recommendations and lay down rules, students and teachers are not waiting. ChatGPT is estimated to have over 100 million users globally and is, by many measures, the fastest spreading digital application of all time, surpassing the vertiginous growth of social media applications, such as Instagram, Snapchat and others.

“Without institutional guidance of any sort, these technologies are likely to get welded into education systems in unplanned ways with uncertain implications and possible unintended consequences. Ideally, there will be serious reflection about their place and role, and then action to realize this vision. We cannot simply ignore the short- and medium-term implications of these technologies for safety, knowledge diversity, equity, and inclusion,” said Mr. Tawil.

91�鶹������Ʒ���� been advising schools and universities to be proactive about providing guidance, and helping learners and teachers better understand these technologies and the implications of their use.

“Educational institutions need an agile and iterative approach, or they will forever be trying to catch up with the relentless pace of technological innovation,” said Mr. Tawil.

Of the educational institutions that reported having a policy, approximately one half said the institution provides ‘pointed guidance’, meaning the institution has clear rules and advice regarding the educational uses of generative AI applications. The other half reported that the institution gives ‘discretion to users’, meaning the institution has largely left it up to individual departments, classes, and teachers to decide whether and how to use generative AI applications.

Of the hundreds of institutions that participated in the survey, only two indicated that they had policies or guidance that constitute ‘a ban’ through which the institution completely or largely prohibits the use of generative AI application such as ChatGPT.

The fact that some 40% of the educational institutions that reported having guidance, said the guidance was not written and had only been communicated orally further illustrates the ad-hoc nature of policy responses in education.  

Universities were significantly more likely to have institutional policies or guidance than schools. Approximately 13% of the universities reported having some sort of guidance, while only 7% of schools did.  

It is also telling that close to 20% of survey respondents reported that they were unsure whether or not their respective institution had policies or guidance concerning generative AI. This significant share reflects the uncertainty and regulatory void currently surrounding these new technologies.

Over the past few years, 91�鶹������Ʒ���� been working to help educational institutions and countries steer the educational use of AI in humanistic directions that prioritize inclusion, equity, diversity, and quality. The 2021 UNESCO Recommendation on the Ethics of AI provides general principles to anchor sector- and country-specific rules and regulations. The 2019 and 2019 publication deal with some of the unique educational and cultural implications of AI technologies, including conversational chatbots. Additionally, the 2021 publication offers pointed policy advice.

The survey was conducted among the global UNESCO networks of Associated Schools and university Chairs from 4 to 19 May 2023. Slightly over 450 institutions responded (11% from Africa, 5% from the Arab States, 23% from Asia and the Pacific, 44% from Europe and North America, and 17% from Latin America and the Caribbean).