The biggest AI flops of 2024

Artificial intelligence has infiltrated almost every corner of the internet

Generative AI makes creating tons of text, images, videos, and other types of content easy. Since it only takes a few seconds between entering a query for your chosen model to spit out a result, these models have become a quick and easy way to produce content at scale. And 2024 was the year we started calling this (generally low-quality) media what it is — AI scum.

This low-stakes way to create an AI stream means it can now be found in almost every corner of the internet: from newsletters in your inbox and books that are resold Amazonto ads and articles across the web and unusual images on your social media feeds. The more emotionally evocative these images are (wounded defenders, the cry of childrenand a support signal in the Israeli-Palestinian conflict), are more likely to be shared, resulting in higher engagement and ad revenue for their savvy creators.

The underutilization of artificial intelligence is not only unpleasant — its rise poses a real problem for the future of the very models that helped produce it. Because these models are trained on data collected from the internet, the increasing number of junk websites containing AI garbage means that there are very real danger models. output and performance will steadily deteriorate.

AI art distorts our expectations of real events

2024 was also the year when the effects of surreal AI imagery began to seep into our real lives. Willy’s Chocolate Experiencewildly informal all-inclusive event inspired by Roald Dahl Charlie and the Chocolate Factorymade headlines around the world in February after its fanciful AI-generated marketing materials gave visitors the impression that it would be far grander than the sparsely decorated warehouse created by its makers.

Similarly, hundreds of people gathered in the streets of Dublin for The Halloween parade that didn’t exist. A Pakistan-based website used artificial intelligence to create a list of events in the city, which was widely shared on social media before October 31. Although the SEO decoy site (myspirithalloween.com) has since been taken down, both events illustrate how the public’s misplaced trust in AI-generated content online can come back to haunt us.

Grok allows users to create images of almost any scenario

The vast majority of major AI image generators have guardrails – rules that define what AI models can and cannot do – to prevent users from creating violent, graphic, illegal and other types of harmful content. Sometimes these guardrails are just meant to ensure that no one is using the intellectual property of others. But Grok, an assistant created by Elon Musk’s artificial intelligence company, called xAI, ignores almost all of these principles in line with Musk’s rejection of what he calls “awakened artificial intelligence.”