There is a type of observational bias in the social sciences called the ¡°streetlight effect¡±:? we only look where it is easy to look.

It¡¯s named for the old joke of the drunk looking for his keys under the streetlight on the sidewalk. When the cop ¨C who had been dutifully helping him ¨C finally asks where he thinks he lost them, the man points into the darkness and replies ¡°in the park¡±. When the cop asks in exasperation why he is looking under the streetlight, the drunk says ¡°it¡¯s a lot easier because this is where the light is.¡±

I feel like that cop when professors say we need to go back to our ¡°old-school¡± ways and bring back blue-book exams, those in-person tests where students handwrite their essays. It took us decades of research to acknowledge that??is a process, and students can¡¯t just produce picture-perfect answers on demand. That¡¯s??we stopped using blue books in the first place: learning had to be student-centred (rather than teacher-driven).

But incredibly, not only are blue books becoming popular again ¨C the 51¹ú²úÊÓƵ of California Student Store??an 80 per cent increase in sales ¨C but some professors even??that ¡°the authenticity and richness of [my] students¡¯ hand-penned prose nearly moved me to tears.¡±

51¹ú²úÊÓƵ

To which I say: ¡°Bah, humbug!¡±

Blue books and handwritten exams more generally are back, of course, because of artificial intelligence (AI). Surveys consistently show that just about all college students are cheating, and professors have almost??for catching them or reversing this trend. Whether it is honour code updates, revisions to academic integrity policies, better AI detectors, AI watermarking, process tracking software, or whatever else, these are all reactive and futile as students play cat-and-mouse games to not get caught.

Blue books are seen as an answer to this problem, and advocates even try to??as more than just a measure of last resort: blue books help us step away from the endless scroll of technology, focus directly on what¡¯s in front of us, and ensure that everything written is indeed from the student. And writing by hand is also??for the brain!

51¹ú²úÊÓƵ

Yet there is so much wrong with this perspective. On a basic level, it minimises and marginalises the 10 per cent of college students with??of learning disability, and it unnecessarily ramps up anxiety for the 10 per cent to 35 per cent of college students for whom such time- and format-delimited stress??their performance. But most problematically, it pretends that we can somehow go back to a transmission model of education, where professors simply transmit knowledge through their lecturing and then grade students on their understanding of such knowledge.

ChatGPT, though,??this model by making the reproduction of polished answers instant and effortless. The old model assumed that what a student could produce under test conditions reflected what they had actually learned; AI severs that link, and no amount of handwriting in a blue book can rewind that reality.

Am I overwhelmed and exhausted from the process of finding a new way of teaching??. Do I wish there was a better way??. But if we want to find our keys (to saving higher education), we need to become courageous and tiptoe into the darkness.

So let me offer some first steps.

First, we have to rethink writing. Since AI has taken away my ability to trust what is and is not students¡¯ authentic final product, I now focus on, grade and scaffold what writing experts have suggested all along: my students¡¯ process of thinking. As the National Commission on Writing??long ago: ¡°If students are to make knowledge their own, they must struggle with the details, wrestle with the facts, and rework raw information and dimly understood concepts into language they can communicate to someone else. In short, if students are to learn, they must write.¡± Writing equals thinking; or to paraphrase Robert Frost more poetically, writing is discovering.

51¹ú²úÊÓƵ

But discovering something is actually really hard. We would all rather skate along on our well-worn??of thoughts and assumptions than have to rethink them. As one recent??put it, thinking is ¡°unpleasant¡±.

And it is really unpleasant (and confounding and disruptive) when we are faced with complex and contested issues, as happens all the time in the college classroom. The second thing faculty therefore have to do is embrace the reality that our job is to guide students¡¯ learning as they tackle such topics. (That¡¯s why it¡¯s called a?!)

None of this demands that we embrace AI; but it also doesn¡¯t mean we have to shun it. AI, when used?, offers powerful scaffolding for everything from brainstorming to clarifying complex arguments and readings. In fact, I work with my college students in literally every class to help them see AI as a conversation partner rather than as a ghostwriter. My goal is always to help my students think and write better. And while it¡¯s a huge amount of work, my students really appreciate having assistance at any time, on any subject, at any level of understanding.

So let me be clear: we don¡¯t need more blue books or a return to in-person exams patrolled by vigilant proctors. The keys to saving higher education won¡¯t be found under the false glow of tradition; they¡¯re somewhere out there in the dark, waiting for us to be brave enough to look.

51¹ú²úÊÓƵ

Dan Sarofian-Butin is professor at the department of education and community studies at Merrimack College in North Andover, Massachusetts.