Harvard University’s introductory coding course, Computer Science 50: Introduction to Computer Science, plans to employ AI as a teaching aid this fall.
A statement by Professor David J. Malan, the course’s director, stated, “Starting in the fall, students will be able to use AI to help them find bugs in their code, give feedback on the design of student programs, explain unfamiliar lines of code or error messages, and answer individual questions.”
CS50: Introduction to Computer Science is an introductory self-taught course that takes 11 to 20 hours per week over 11 weeks. While the course is internationally recognized, it’s free to participate. Approximately 5 million people are currently enrolled in the course.
Student use of generative AI like ChatGPT is becoming harder for educational establishments to avoid.
Some establishments have lept to ban ChatGPT altogether, while others are taking a more cautious approach. UK Russell Group universities recently issued a joint statement on AI with 5 principles for harnessing the technology while preserving academic integrity.
One thing is for sure – AI will transform education. Speaking in Japan, OpenAI CEO Sam Altman compared AI to the calculator, whereas Nvidia CEO Jensen Huang compared going to university now as similar to going to university when the personal computer (PC) became widespread.
The debate reflects narratives surrounding AI’s societal impacts – balancing risks and benefits is exceptionally complex and requires collaboration between AI developers and users.
How Harvard intends to use AI
Malan highlighted CS50’s software-integrated approach and labeled this as an “evolution of that tradition.” He explained, “Our own hope is that, through AI, we can eventually approximate a 1:1 teacher:student ratio for every student in CS50, as by providing them with software-based tools that, 24/7, can support their learning at a pace and in a style that works best for them individually.”
To assist students throughout the course, the “CS50 bot” will answer student questions on Harvard’s STEM discussion board. Human staff can review the AI-generated responses.
However, Malan observed that AIs such as ChatGPT and GitHub Copilot are currently “too helpful” – or, in other words, too good at programming. Therefore, “In CS50, the AI technology will be ‘similar in spirit’ but will be leading students toward an answer rather than handing it to them.”
The AI incorporation in CS50 will also extend to its online counterpart on edX, a platform developed through a collaboration between MIT and Harvard. Malan wrote, “Providing support that’s tailored to students’ specific questions has long been a challenge at scale via edX and OpenCourseWare more generally, with so many students online, so these features will benefit students both on campus and off.”
Of course, there have been criticisms. In an interview with Fox News, Martin Rand, PactumAI co-founder and CEO, warned, “I would say the dangers are that we have to consider that these are statistical models. These will come up with most probable answers and high probability can also mean mediocrity. So professors need to be there to provide exceptionalism, and I think Harvard has taken the right approach in providing this only to introductory courses.”
Malan addressed some of these points, saying, “We’ll make clear to students that they should always think critically when taking in information as input, be it from humans or software.”
AI’s role in education should be monitored closely to avoid risks such as enfeeblement, where humans become overly reliant on the technology instead of applying their own skills and intelligence.