ChatGPT is really bad at procedural tasks

This image features a close-up view of an airplane safety instruction card. The card is divided into multiple panels, each containing illustrations and instructions for various safety procedures. These include how to use an oxygen mask, how to inflate a life vest, and the correct brace positions in case of an emergency. The illustrations are colorful and designed to be easily understandable, using a combination of images and minimal text to communicate critical safety information. The overall layout is designed for quick reference, catering to passengers' needs during an emergency.

In this post, I look at ChatGPT’s ability to create procedural tasks in images. Using paper airplanes, three-fingered dwarves, origami rabbits, and the sinking Titanic we look at the limitations of this powerful tool in creating coherent instructional images.

AI image generation: From text to images

A digital image displaying a matrix of green alphanumeric characters cascading vertically on a black background, reminiscent of digital data streams or the iconic 'Matrix' code rain.

Generative AI has been filling our newsfeeds and dominating discussions over the past 18 months. In this post, I explore the feelings I have toward AI image generation and the opportunities we have to teach information consumers about this powerful new generative AI technology.

Blurring realities: Indistinguishable human and machine creations

A digital art piece showcasing the convergence of traditional higher education and futuristic AI. It depicts a person stepping from an open book, symbolizing conventional learning, through a luminous, code-lined portal into a digital space. Classic university buildings flank the portal, while above, digital and educational elements merge, with neural networks and technological devices floating in the sky. This represents the theme 'Entering the Matrix: When Human and Machine Creations Become Indistinguishable' for a blog post header.

Explore the nuances of AI integration in higher education and the digital realm, highlighting the necessity for conscious engagement with indistinguishable AI involvement. Addressing the blend of AI-generated and human content, the author advocates for visual cues to make AI’s role discernible, underscoring transparency in an increasingly machine-generated world.

Inside my tech bag: Unveiling the power of RODE Wireless GO II

A stylized blog header image depicting a vintage library setting with wood bookshelves full of books. In the center, a modern, detailed illustration of a microphone (Yeti styled) sits atop an open book, symbolizing the fusion of traditional learning and contemporary technology. Surrounding the microphone are various icons and doodles representing audio and recording equipment, signifying the microphone's utility in different scenarios. The image has a decorative border reminiscent of art deco style, with vertical bars on the sides that suggest sound level indicators. The color palette consists of deep wood tones, black and gray for the microphone, and light blue accents, creating a professional and academic atmosphere.

Discover how the RODE Wireless GO II isn’t just a gadget—it’s a revolution in teaching technology. Compact yet powerful, this microphone system transforms any space into a stage for clear, professional audio capture, enhancing my presentations and ensuring no student misses out, no matter where they are.

The debate around ChatGPT in academic peer-review

An artistic blog header image depicting a conceptual blend of classical and digital learning. The left side of the image presents a traditional library with rows of bookshelves and an open book on a desk, lit by warm, glowing lights. Transitioning to the right, the scene morphs into a digital realm with several laptops open, displaying futuristic, holographic data and codes, illuminated by cool blue light that emanates from a central, bright source at the end of the library's aisle. The image represents the merging of time-honored academic research with modern artificial intelligence technologies.

Explore the growing concerns and debates around ChatGPT’s role in academic peer-review processes. This insightful piece delves into real-world examples of AI-generated content being submitted without attribution, raising questions about the effectiveness of current review systems and calling for heightened vigilance to preserve the integrity of academic publishing.

ChatGPT: Clever, but not infallible

As I explored the capabilities of ChatGPT, I put it to the test by tasking it with generating an annotated bibliography entry for a well-known work, Paulo Freire’s “Pedagogy of the Oppressed“. To my surprise, it delivered a satisfactory entry in just 20 seconds. However, I decided to challenge ChatGPT … Read more