Letting AI do the soiled work of programming frees college students to work on problem-solving. Issarawat Tattong/Second by way of Getty Photos
Unusual Programs is an occasional sequence from The Dialog U.S. highlighting unconventional approaches to educating.
Title in fact:
“Be taught AI-Assisted Python Programming”
What prompted the thought for the course?
Generative AI is de facto good at pc programming – to the purpose the place the way in which we educate and assess college students who’re studying to program should change.
We used to present college students dozens or a whole bunch of small focused programming duties, drilling every side of the syntax – the phrases and symbols – of programming. That labored nicely as a place to begin, besides now generative AI instruments can clear up all of those issues. Educators can attempt to ban these instruments (good luck with that!), or embrace them. We selected to embrace them in our new course, the place college students study to program – supported by a generative AI assistant.
What does the course discover?
The course re-imagines what studying to program means now that generative AI is offered to deal with extra of the low-level syntax points which have traditionally slowed down and annoyed college students. The extra college students wrestle with finicky syntax particulars, the much less time and vitality they’ve to perform their programming-related objectives like beginning a enterprise, writing apps for social good, or contributing to initiatives which are significant to them.
Generative AI clears the decks for us to give attention to extra helpful, high-level abilities that college students have to change into efficient programmers. For instance, generative AI struggles to unravel giant issues; we nonetheless want people to divide these issues into bite-sized chunks – a course of referred to as drawback decomposition – every of which AI can clear up nicely. Individuals are nonetheless wanted to check code to make sure it’s doing what was supposed, and to make sure that the code is used to assist, not hurt, society and its weak teams.
Why is that this course related now?
Skilled programmers in droves have already adopted generative AI instruments and are utilizing them to be extra environment friendly of their day by day work. If the aim is to arrange college students for these jobs, academics want to coach them in the way to use these new instruments.
Maybe extra importantly, what college students can do in introductory programs modifications. With a extra highly effective instrument comes a capability to work at greater, extra environment friendly ranges. These instruments save individuals time.
AI code assistants are altering what it means to do pc programming.
What’s a essential lesson from the course?
A essential lesson is that generative AI is spectacular, however that it’s fallible. You can not merely ask it for code and assume that the code it provides you is ideal. It could not do the proper factor. It could produce errors, or bugs. It could trigger safety considerations. It could exclude underrepresented teams or discourses. It’s essential to critically look at the code that you simply get from generative AI.
What supplies does the course function?
The course is constructed on our new e book “Be taught AI-Assisted Python Programming.” The e book reconceptualizes an introductory programming course within the context of generative AI instruments.
The principle instrument used within the e book and in our course known as GitHub Copilot, which is like ChatGPT for programmers. College students use Copilot from day one. They construct full apps: apps to automate tedious, error-prone duties; pc video games; even an app to guess who wrote a novel whose creator could also be unknown. To make sure that college students are nonetheless studying fundamentals, the e book teaches them the way to perceive the code that the generative AI is creating.
What is going to the course put together college students to do?
Some college students take an intro programming course to begin their pc science main. For these college students, we proceed to show evergreen abilities like code studying and code testing, however now additionally introduce the higher-level ability of drawback decomposition so college students can clear up bigger duties than ever earlier than.
Nearly all of college students within the course, although, are learning different disciplines like sociology, psychology, enterprise, engineering and science. The course prepares these college students to make use of generative AI to spice up their careers by way of programming.
Leo Porter receives compensation for gross sales of the e book "Be taught AI-Assisted Python Programming."
Daniel Zingaro receives compensation for gross sales of the e book "Be taught AI-Assisted Python Programming." He additionally consults on books for Manning Publications.
