Quandaries of the modern professor

Two challenges strike me this week:

1. What to do about student use of artificial intelligence (AI) tools to prepare written product.
2. What to do about the risk that a student may be a spy for a foreign government.

These are real-life quandaries.

AI tools in the classroom

ChatGPT and other AI tools that prepare written product in response to user inquiries are the immediate problem. They go a step beyond what we are all used to. We all use search engines to look for relevant bibliography. We also read Wikipedia to get oriented to a new subject or check facts. I am fortunately not teaching classes with assigned written products these days, but I am supervising graduate students preparing theses. What do I think about use of generative tools to produce written material? Do I feel the same about their use as an exploratory tool and as a final product?

It seems to me the answer to that second question has to be no. Final product has to come from the student her/himself, with citations to sources. But I can’t really object to AI use as an exploratory tool. We all need to start someplace. My own habit is to start as much as possible with primary source materials, but I am an historian. I don’t really see why a political scientist or an economist shouldn’t ask generative AI to summarize the latest data on election results or business cycles. That is a small step past using a search engine on those terms. Even if current AI technology doesn’t deliver much, no doubt future generative tools will.

Learning is key

The real issue then is how you determine what the student has learned, beyond the products of the AI tools s/he may have used. It seems to me that isn’t so hard. Exams (without computers) can serve that purpose. So too can oral presentations or informal chats with the professor.

When I was a graduate student preceptor (teaching assistant) at Princeton, I received a paper from a student who hadn’t performed well in class. He referred in the first paragraph to the “pre-lapsarian Adam.” When I asked him what that meant, he was unable to respond. Plagiarism is not new. I didn’t need “Turnitin,” a software that can now check for it.

Spies in the classroom

No one who teaches international affairs anywhere on earth can be sure he/she hasn’t had a spy–past, present or future–in the classroom. Certainly a goodly number of students at SAIS, both Americans and non-Americans–later pursue careers in their home country intelligence agencies. We also get students who pursue degrees while they are still working for government agencies, or after they have completed careers in them, with intelligence responsibilities. Teaching them analytical methods and policy frameworks is a good idea, not a bad one.

The problem is current spies who hide their true identities. That is what the Washington Post says SAIS’s Russian student pretending to be a Brazilian did. The Post focuses on what he himself might have been doing. The example they give is reporting last year on US attitudes towards a Russian invasion of Ukraine. That Moscow could have gotten reading the daily press. The Post also focuses on what he might have done in the future. Penetrating the International Criminal Court’s computer systems is the case in point, as he had an internship lined up there before being arrested. That seems to me more serious.

Another issue

But I would point to another issue: recruitment. International affairs schools are brimming with students who will go on to bigger and better things, in government and in the private sector. A covert agent might have a field day lining up people living on student stipends and finding it difficult to pay tuition or otherwise make ends meet. Once recruited, such an individual becomes subject to lifetime blackmail.

The consequences could be long term and catastrophic.

The tougher problem

Spies in the classroom seems to me a much tougher problem than AI in the classroom. A professor can be expected to know when students are learning. But professors don’t command the tools required to ferret out covert operatives. That is an intelligence and law enforcement responsibility, not an academic one. Covert spies in the classroom could compromise classmates (or professors) and create problems for decades in the future.

There are more than 750,000 foreign students in the US. Do we really think the CIA and FBI are capable of keeping an eye on even a small fraction of them? And would we want the limitations that would necessarily come if they tried to do so?

No easy answers

Nor would I want the university administration to take on that responsibility. The most we can expect from it is to make reasonable efforts to ensure students are not using false identities. Beyond that, I suppose we’ll need to keep a watch out for suspicious behavior. But I confess as a Foreign Service officer I met several Americans who spied for foreign governments: Walter Kendall Myers (Cuba), Aldrich Ames (Soviet Union and Russia), and Felix Bloch (Soviet Union). I suspected none of them.