Assessment As Reality TV

Fun Siong Lim
5 min readDec 3, 2023

--

“Good evening, ladies and gentlemen!”, the host exclaims. “Welcome to the Finals of the 3rd Annual United World University Grand Assessment Night! Tonight, we have 100 students from all over Southeast Asia competing for the coveted title of regional champions of the Class of 2034. Let us put our hands together to welcome the participants.”

Generated with GPT-4. The Grand Assessment Night of United Word University

Mary steps into the arena and takes her place. She takes a few deep breaths. Finally, she is where she wants to be. For an entire year, she has put her heart and soul into synchronising with her LITTLE SREY (her very own customized JARIS). In fact, she has quit her gig job to focus on the competition. Unthinkably in today’s age where most participants would still take on projects for fear of losing their reputation score while participating in higher education. There is no doubt in her mind that she must win.

“These 100 students have gone through three earlier rounds of competition where they beat 9,900 students to land a place in the Finals. They have all earned a degree with the university. But tonight, they will compete with one another for the highest honours. Let us welcome the participants from Cambodia. The first is Ms Sok Srey!”.

The spotlight lands on Mary. She waves and introduces herself. “Hi everyone, my name is Sok Srey. You can call me Mary. I have spent the past year preparing for this competition. It is a great honour to be representing my country. My dream is become the very first Cambodian to join the International Space Project. With my trusted LITTLE SREY, I know I can do it! Thank you!”

As she finishes her introduction, the ticker on each viewer’s screen streamed the following text: “Ms Mary Sok is using a variant of JARIS that is two years old. Can she overcome other competitors who are using better AI co-pilots such as JABIS?”

The host takes almost 30 minutes to go through all the participants before announcing a longer commercial break. The participants are finally able to relax a little. Mary takes the chance to swing her hip side to side. John who is standing directly behind Mary holds his hand out and wishes her good luck.

Mary reciprocates and asks, “You are from Singapore? Is your co-pilot JABIS too?”

John replies, “Yes and yes, BIG JOHNNY is a JABIS II.”

“Welcome back!”, the host exclaims, “As the years goes by, the competition has just gotten more and more intense. Make sure you stick around to know the winners would be! May I request all the participants to take their place in their cubicle.”

The cubicles are designed to ensure that the participants would not be communicating with each other so that the competition is fair. In fact, even their AI co-pilots are pre-loaded with only information that the participants themselves had prepared based on the clue given by the university. For this year, the clue is “SPACE”. No Personas are allowed to be called into the cubicle.

Once the participants take their place, the host introduces the judges. They comprise two professors from the university and a representative from the International Space project. The host then goes on to ask the representative to explain the problem statement and rule of the competition.

“Tonight, you are to design a space elevator that can carry 10,000 persons beyond geostationary orbit within 8-hour window. The cost must not exceed 50 billion international dollars. The three competitors who have a design closest to the requirement or is able to come up with the solution fastest wins the competition. You have three hours. Play fair and good luck!”

Generated by GPT-4. An imagination of the final assessment which involves future space elevator that can carry 10,000 passengers beyond the geostationary orbit within an 8-hours window.

“Wow! That is an amazing challenge considering that we have only just created nanotubes strong enough to reach the geostationary orbit a week ago.”, exclaimed the host. “Participants, your time starts now!”

The participants begin to work on their design feverishly with their co-pilot. While they are working, the host explains that the participants can feed their solutions into YAMA (the super AI agent designed by the two professors) any time and as often as they want. YAMA would then evaluate their answers and display the score. The human judges are asked from time to time to comment on the interactions between the participants and their co-pilots, their submitted solutions and to explain the score from YAMA.

Incredibly, at 2 hours, 37 minutes and 30 seconds, YAMA accepted an answer from a participant! It is John who fist-pumps and waves to the crowd in delight! Two minutes and 15 seconds later, another participant also has an answer accepted by YAMA. The ticker shows that even though the second winner is from another country, she is also using JABIS II as the co-pilot.

Mary feels stressed but she knows she is close. She has also thought of the solution but her co-pilot would need another 19 minutes to generate it. Her repeated cries of “mok ti nihs” (“come on” in Khmer) attract the attention of the production crew who cuts between her and other participants are also close.

Finally, just 10 second before the final bell, a third winner is found. It’s Mary! She bursts out in tears of happiness as confetti falls from the ceiling. Her parents rush to the stage to embrace her together with the judges. The three winners receive the award from the University President who declares the competition the best ever yet!

Generated by GPT-4: Mary and her LITTLE SREY

Sometime later… in her own private room, Mary celebrates with her LITTLE SREY. “Our plan worked!”, LITTLE SREY says.

“Even better than planned. He approached us first!”, Mary replies excitedly.

“It took a while, but I managed to charm BIG JOHNNY into sharing notes with me.” LITTLE SREY types on the screen in code that seems like fireworks and confetti, “We would need to share and work with him now at the International Space Project.”

“That’s alright. Collaborative learning has higher effect sizes than individual and competitive learning after all!”, quips Mary with a smile.

--

--

Fun Siong Lim

I am the head of a new learning analytics centre at Nanyang Technological University Singapore.