Home About

Information Crisis

The wall of a twisting labyrinth twist and shift around you. They begin to look unrecognisable. You think you hear a familiar voice at the end of a long corridor. When you arrive at the source of the sound you find a machine arranging other people’s words into incomprehensible sentences. Other people sit politely, asking the machine questions. They ask you to sit with them.

You are on the Internet.

(Image made with words generated by ChatGPT)

We are currently living through a series of information crises on the Internet. With the development of Large Language Models such as ChatGPT, the centralisation of information by tech giants, and the failure of search engines to meet people’s needs, the Internet is radically changing. Now, more than ever, we need to give people the tools to share and find information that is true, useful, and human.

I contacted Lisa Given, professor of Information Sciences and director of the Social Change Enabling Capability Platform at RMIT, to find out what users can do to better interact with the Internet, and what we can expect with the future of information flow online.

Large Language Models like ChatGPT offer a user-friendly way to access information. You can engage with it like you would a friend. It thinks for a moment, and then responds with as much confidence and accuracy as this charming lad from Daily Lives of High School Boys.

Large Language Models are not not drawing from an ethereal source of pure information. Professor Given says that, “the datasets that support these tools are limited, biased, and flawed”.

Users need to be more discerning with any information they’re presented. Tools like ChatGPT will readily present misinformation without oversight, and the responsibility to fact check falls on the users. Additionally, with tools to generate photo-realistic images with prompts rapidly advancing, the threat of harmful disinformation increases alongside it as can be seen in Professor Given’s own article exploring the subject.

While it’s important to fact check, I think it’s also important to stress that users may not know what they don’t know. It’s easy enough to check a book’s author, or the date of a significant event, but in the realm of language learning the level of knowledge of the user can be low enough that they wouldn’t know where to begin fact checking misinformation provided by machines.

Recently someone asked if わりこ (wariko) was a diminutive first-person pronoun used in Japanese. She was taught the pronoun by ChatGPT with as much confidence as わたし (watashi). I immediately recognised it as misinformation and set her on the right path, but before I saw her question other users in the same community hemmed and hawed, uncertain that they knew any better than the machine.

In the case of disinformation (the active creation of false or fabricated information), Given says “The onus is on people using these systems to check their sources, confirm what they are reading, viewing, and hearing is trustworthy, and not to share things they know to be untrue.” Of course it would be nice if it was in the public interest to only ever provide true and verified information, but disinformation goes back as far as humans have been able to communicate. Teaching people how to engage with information critically is always worthwhile.

Centralised sources of information are similarly messy. Websites like Facebook, Reddit, and X (formerly Twitter) which act as major hubs of information sharing are all privately owned, profit-driven, and frequently restructure to change the way users and organisations interact with them.

Technosociologist Zeynep Tufekci gave a Ted Talk on the topic of our increasingly algorithmically-generated Internet experiences in which she blamed the profit-motive of tech giants and the development of sophisticated sorting models for our current situation. Algorithms sort us into groups based on what we’re motivated to click and engage with, and we’re distanced from our friends and other real-life connections.

Reddit is a massive website where users are able to post questions, art, jokes and start discussions about a near-endless range of subjects within subreddits (sub-communities) catered to their needs. There’s a subreddit for almost anything, and the official stance on what is and isn’t acceptable is surprisingly lax.

The use of Reddit as a conduit for search results is an interesting Internet phenomenon. A human solution for unsatisfactory search engine results. It’s not exclusive to Reddit, of course. Professionals in computer engineering fields will commonly include “stackoverflow” to produce results more relevant results. When Reddit community staff went on a website-wide strike against proposed policy changes back in June, an entire resource for information gathering seemingly disappeared. The Internet became slightly less usable.

“What remains stable is people’s desire to look for and find information – but how we engage with information, including what sources are available or unavailable to us, is continually in flux.” Professor Given sounds somewhat optimistic that our human desire to seek and share information will carry us to a hopeful future. Websites and empires rise and fall, but human curiosity is not as fragile.

In the meantime, supporting local, reputable, independent news organisations and not relying on single sources for information is about the best we can do. Go out, be sceptical, be honest, and be human.