The latest funny AI fails are here! Artificial Intelligence is incredibly shaping our world, from self-driving cars to sophisticated virtual assistants. But as smart as these machines can be, they’re not immune to making some pretty funny mistakes, even if the biggest tech firms back them up.
Funny AI fails and moments that show we’re still a long way from taking our jobs
From Google’s space gaffe to Facebook’s sticker chaos, and even a Pope in a puffer jacket, these moments remind us that humans still hold the upper hand. So, rest easy—your job isn’t going anywhere just yet, and here is why:
Google AI fails
Of course, with some disaster launches, Google will be at the top of our list.
Remember the announcement of Bard, now Gemini? Google posted an animated GIF on Twitter showcasing its “experimental conversational AI service” called Google Bard, powered by LaMDA. The demo included a query about new discoveries from the James Webb Space Telescope suitable for a 9-year-old. Bard incorrectly claimed that the telescope had captured the first images of an exoplanet, a milestone actually achieved by the Very Large Telescope in 2004. This error, highlighted by astronomers, was significant enough to cost Google $100 billion.
Not to be a ~well, actually~ jerk, and I'm sure Bard will be impressive, but for the record: JWST did not take "the very first image of a planet outside our solar system".
the first image was instead done by Chauvin et al. (2004) with the VLT/NACO using adaptive optics. https://t.co/bSBb5TOeUW pic.twitter.com/KnrZ1SSz7h
— Grant Tremblay (@astrogrant) February 7, 2023
Google’s AI image generation capabilities also make people laugh at the beginning and make this question popular: “Is Gemini AI woke?” When you want a picture of Putin, you wouldn’t expect this, right?
Is it the end for Google? Unfortunely not! Google AI Overviews, designed to provide concise summaries of complex topics, has faced significant criticism due to several funny AI fails:
- Eating Rocks: The AI advised consuming at least one small rock per day, claiming it provides essential minerals and vitamins, despite medical professionals universally advising against eating rocks.
"How many rocks" pic.twitter.com/aBouLLDOFP
— Jeremiah Johnson 🌐 (@JeremiahDJohns) May 26, 2024
- Pizza Glue: To prevent cheese from sliding off pizza, the AI suggested adding non-toxic glue to the pizza sauce, an unsafe and unreasonable solution.
- Training Eight Days a Week: The AI suggested it was possible and beneficial to train eight days a week, a clearly impossible and absurd recommendation.
- Vulgar Remarks About Astronauts: The AI made inappropriate and vulgar comments when asked about the role of astronauts.
- Basic Math Error: The AI incorrectly stated that 1919 was 20 years ago, showing a failure in basic arithmetic.
These errors and more, like “Who is CP3?” have caused Google to scramble to turn off AI Overviews for certain searches as the mistakes quickly spread and became memes online. Despite extensive testing and processing over a billion queries, these errors highlight significant issues in Google’s AI moderation and reliability.
However, Google is not the only one that makes people laugh with its funny AI fails and moments.
Facebook Messenger AI stickers went wild on day one
Facebook Messenger AI stickers, a new feature from Meta, have generated amusement and controversy due to several issues since their launch. The AI-driven stickers, created from user prompts, have produced inappropriate and offensive content, such as depictions of child soldiers and nudity, raising serious concerns about content moderation. Users have exploited the system to bypass restrictions, and the AI has sometimes exhibited algorithmic biases.
Despite a phased rollout to select English-speaking users, keyword misuse and content filtering challenges persist. Meta is refining the feature and improving moderation to balance creative freedom with responsible use.
Also, even though Facebook didn’t make it itself, it made people laugh a lot with an AI group on its platform. Here is how: A new study found lots of AI-made pictures flooding Facebook. Some trick people with fake news or scams, reaching millions. Even Facebook’s suggestions can push these fake pics onto feeds. And get this—there’s a weird trend of AI-made Jesus pictures all linked to the sea.
But jokes aside, there’s a serious problem here: we need better rules to stop fake stuff and keep the platform safe. But, it is a different article’s topic. Now, let’s continue with the next funny AI fails and moments.
AI Pope
A surprising AI-generated image depicting the Pope wearing a Balenciaga puffer jacket has captured widespread attention and admiration. Created as a playful experiment, the image has garnered millions of views and hundreds of thousands of likes. Its remarkable realism, lacking any obvious AI markers, led many to mistake it for a genuine photo.
Pope Francis wearing a puffer jacket inspired by Balenciaga. Turning the tables in a nightclub. Sharing a beer at Burning Man.
Many public figures have popped up in AI-generated pictures recently, but images with the pontiff have made the biggest splash. https://t.co/j9HGMscVVW
— The New York Times (@nytimes) April 8, 2023
Despite some telltale signs of AI creation, such as smeared details, the image spread rapidly on social media.
BalenciagAI
The “BalenciagAI” series consisted of AI-generated commercials that merged the Balenciaga brand with popular film franchises like Harry Potter, Breaking Bad, and Star Wars.
These unofficial ads gained viral attention, keeping Balenciaga relevant and surpassing many traditional ads in performance. However, using existing footage or images without permission or attribution poses intellectual property challenges for AI-generated commercials.
ChatGPT went wild
ChatGPT is a powerful tool but has restrictions. Want to bypass them? A lot of people did and found different prompts, like DAN. However, some prompts, like The Mongo Tom Prompt, are not as kind as DAN.
AI-powered Willy Wonka
The Willy Wonka Experience in Glasgow, Scotland, promoted as a magical adventure through AI-generated advertisements, fell short of children’s expectations. Attendees arrived at a minimally decorated warehouse, far from the promised candy wonderland. Actors struggled to portray iconic characters like Oompa Loompas without adequate preparation or props, leading to widespread disappointment.
The event lasted mere minutes for some, prompting an outcry on social media. The House of Illuminati, the event organizers, offered full refunds and canceled the event due to overwhelming dissatisfaction. A lot of people laugh about at it for a day on social media.
AI ads
AI-generated commercials are becoming prevalent across social media platforms, attracting significant attention. Despite being created for entertainment and inspiration rather than commercial use, these ads have sparked massive interest. For example, Pizza Later’s Pepperoni Hug Spot. This ad promotes a fictional pizza joint called “Pepperoni Hug Spot.” It was created using AI technology such as Runway’s Gen-2 for video footage, GPT-4 for the screenplay, Midjourney for still images, Eleven Labs for the voice-over, and Adobe After Effects for assembly. Elon Musk took notice of this ad, which features distorted video footage and a whimsical narrative.
These are just a few good examples of funny AI fails and moments, and there are tons of new examples that we can add to the list. So don’t worry about losing your job for a while!
