The AI Conundrum in our Day Schools and Yeshivot: In Search for Best Practices
Given the recent use and application of Artificial Intelligence in our Jewish day schools and yeshivot, the topic of AI best practices is a relatively new challenge, which is moving at lightning speed.
Not unlike best practices in other fields, the more we learn about AI, the more critical it becomes that we identify and create protocols for its use and application in education.
This reality is essential in light of its potential use or misuse in our schools, community and in our homes. In fact, never before in recent memory can I recall such discussion, concern and debate regarding the potential impact of technology on our lives as with AI.
One of the reasons for this concern may be due in part to its unlimited potential power and how much control we may have (or not have) over its use and its powerful, application and impact as a formidable alternative to human cognition and intelligence.
In Our Schools
As we know, the ubiquitous nature of AI technology may eventually make it very challenging to control or suppress. It is therefore imperative that our schools develop best practice protocols and policy guidelines for its use
The following are best practice suggestions for AI use in our schools.
- Before a school embarks on any use of AI, it must first develop an internal infrastructure with appropriate filters. Devices such as Chromebooks managed by Google Admin suggests that no personal laptops be permitted on the school’s Wi-Fi and that all AI sites must be blocked by default then whitelisted only;
- Schools must install a filter that filters AI itself.
- School WI-Fi must have two distinct networks. One for staff and administration and a separate one for students.
- It is essential that all day school and yeshiva staff, irrespective of position, participate in ongoing AI technology training relating to its use, misuse, utility and capacity;
5. Students should never be permitted to use AI on campus unless it’s through filtered school Chromebooks;
6. Students must not be permitted to use image generators, even filtered ones. It will invariable create major challenges regarding inappropriate viewing and tziniut issues;
7. Don’t ever use free Chatgpt with students;
8. Never skip filter updates;
9. Always restrict Chromebook assignments to in-classroom use as opposed to allowing students to use unsupervised technology at home; and
10. Never allow students to bring their own personal devices to school, no matter how much pressure there is coming from parents.
The Home
The home is probably where 90% of the problems actually happen. Schools can lock down Chromebooks, but the home is a wide open environment. If parents are not provided with a clear plan, their children will learn their AI habits from Tik Tok or from other inappropriate platforms. The risks are profound and the consequences are significant.
Several policies regarding AI use in the home may include:
- Permitting students to only used AI at home that is provided and monitored by the school;
- Parents and students must sign contracts regarding the appropriate use of AI technology. Contracts which are broken may lead to detention or permanent dismissal from school;
3. AI use for homework at home must be supervised by parents and it must be on a shared family device in a common area of the home, not in a bedroom or behind a closed door.
4. No AI use for schoolwork after 9:00 pm for elementary and 10:00 pm for high school students;
5. All parents must attend a series of mandatory parent workshops and seminars regarding the proper use of AI technology for their children;
6. If a student is caught cheating through the use of AI, the school will enforce serious consequences leading to possible expulsion from school;
7. The use of “Character AI” is absolutely forbidden; and
8. All homework must be AI proof. Parents do not have to police cheating because the assignment design will make cheating obvious.
For Teachers
- Never paste a student’s full name, grade or IEP into free AI;
- Never ever use AI to give psak halacha or hashkafic advise to students;
- Never use AI to general images of students or people;
- Never grade a test, or quiz with AI. You are the teacher or Rebbe, not the bot;
- Never ever assume that AI is correct. Always check (and double-check) primary sources.
Given the complexity of AI use and misuse in education, there are literally volumes written about AI school policies, best practices and guardrails which must be carefully enforced.
In order to more fully address these concerns responsibly, the school administration may consider establishing a working professional Task Force that is charged or mandated with developing strict policies as warranted. A school that neglects this responsibility is a school that places their students in great jeopardy or harm
Other AI Policy Considerations and Assumptions
If a student always uses AI to write, he will probably never learn to think clearly, structure an argument or find his own voice,
From a Judaic Studies perspective, memory and yedios are critical. As such, if you can always ask AI, why memorize chumash, mishnayot, or historical dates. Moreover, AI does not know truth. It predicts the next most likely word.
For example, AI will invent sources, meforshim and gemaras that may not necessarily exist, give a wrong answer with total confidence and/or a citation that looks ideal.
AI does not know the truth. It predicts the next most logical word. That means it can give the wrong answer with total confidence and a citation that looks real and authentic.
In a Judaic context, this can be very dangerous. A student asks “what does the Rambam say about X” and gets a beautiful and completely fabricated answer. It a student dos not check a real source, they will not catch it. Moreover, AI is not a posek. It has no mesorah, hashkafa or yiras shamayim.
Artificial intelligence will provide you with a “Jewish perspective” that represents a blend of everything with no loyalty to the school’s derech or hashkafa,.
Remember, it will never replace the rebbe or teacher relationship.
Finally it is important to note that students with a more refined command of the English language, and savvier parents, get much more out of AI than others.
Although there are a wide variety of reasons why AI is not 100% accurate, over a period of time, as the technology develops with more comprehensive algorithms, its responses will become more accurate.
The challenge for the user is to determine when that will actually occur.
At the end of the day, educators, teachers and schools must be extremely diligent and careful how and when they us AI.
As discussed, the technology can serve as a most valuable source and resource, tool or it can take you down a path fraught with errors and mistakes.
Until AI is refined, the educational community must be put on notice that the technology is still in its infancy. Having said that, as AI technology evolves and advances at meteoric speed, we will be challenged to ensure its proper and responsible use. This includes not being dependent or over dependent on what has now become one of the most significant technological advances of the 21st century.