🕷️ Crawler Inspector

[Search](https://www.popularmechanics.com/search/)  *** - [Newsletter](https://popularmechanics.com/email/popularmechanics-newsletter) - [Pop Mech Pro](https://www.popularmechanics.com/about/a40180523/pop-mech-pro-membership-news-benefits/) - [Subscribe](https://shop.popularmechanics.com/popular-mechanics-all-access-3499.html?cds_tracking_code=poppanelnav) - [Home and DIY](https://www.popularmechanics.com/home/) - [Power Tools](https://www.popularmechanics.com/home/tools/) - [Cars](https://www.popularmechanics.com/cars/) - [Technology](https://www.popularmechanics.com/technology/) - [Military](https://www.popularmechanics.com/military/) - [Science](https://www.popularmechanics.com/science/) - [Adventure](https://www.popularmechanics.com/adventure/) - [Sci-Fi](https://www.popularmechanics.com/culture/) - [Airplanes](https://www.popularmechanics.com/flight/) - [Space](https://www.popularmechanics.com/space/) - [Gifts](https://www.popularmechanics.com/holiday-gift-guide/) - [Deals](https://www.popularmechanics.com/sales-and-deals/) - [Visual Stories](https://www.popularmechanics.com/interactive/stories/) - [Promotions](https://www.popularmechanics.com/promotions/) - [Shop](https://shop.popularmechanics.com/?source=pop_hamburger) - [Follow](https://www.popularmechanics.com/technology/security/a43013341/why-ai-doesnt-understand-memes/#follow) - [My Bookmarks](https://www.popularmechanics.com/my-stuff/) [Privacy Notice](https://www.hearst.com/-/us-magazines-privacy-notice)[Terms Of Use](https://www.hearst.com/-/us-magazines-terms-of-use) [Skip to Content](https://www.popularmechanics.com/technology/security/a43013341/why-ai-doesnt-understand-memes/#main-content) [Newsletter](https://www.popularmechanics.com/email/popularmechanics-newsletter) [](https://www.popularmechanics.com/) [sign in](https://www.popularmechanics.com/auth/csrf?action=login&return_url=https%3A%2F%2Fwww.popularmechanics.com%2Ftechnology%2Fsecurity%2Fa43013341%2Fwhy-ai-doesnt-understand-memes%2F) [Subscribe](https://shop.popularmechanics.com/popular-mechanics-all-access-3499.html?cds_tracking_code=poptopnav) [](https://www.popularmechanics.com/search/) - [Science](https://www.popularmechanics.com/science/) - [Home & DIY](https://www.popularmechanics.com/home/) - [Technology](https://www.popularmechanics.com/technology/) - [Gear Reviews](https://www.popularmechanics.com/product-reviews/) - [Military](https://www.popularmechanics.com/military/) - [Transportation](https://www.popularmechanics.com/transportation/) - [Culture & Curiosities](https://www.popularmechanics.com/culture/) - [About](https://www.popularmechanics.com/about/a45/about-us/) - [More](https://www.popularmechanics.com/technology/security/a43013341/why-ai-doesnt-understand-memes/)  Popular Mechanics; Getty Images 1. [New Technology](https://www.popularmechanics.com/technology/) 2. [Security](https://www.popularmechanics.com/technology/security/) 3. [Artificial Intelligence Doesn’t ‘Get’ Memes. Here’s Why That Could Save Us From the Singularity](https://www.popularmechanics.com/technology/security/a43013341/why-ai-doesnt-understand-memes/) # Artificial Intelligence Doesn’t ‘Get’ Memes. Here’s Why That Could Save Us From the Singularity Our silly little jokes totally confuse AI, which proves “humans have the upper hand over machines,” an expert tells us. *At least, for now.* By [Matt Crisara](https://www.popularmechanics.com/author/228667/matt-crisara/ "Matt Crisara") Updated: Jan 29, 2024 2:42 PM EST Save Article  1 Listen (11 min)11 min Artificial intelligence is supposedly drawing closer to [singularity](https://www.popularmechanics.com/technology/a42828175/what-is-the-singularity/)—the point at which machine intelligence surpasses human intelligence. That’s sparked debates about how we can best control AI’s capabilities; while superintelligent systems *could* revolutionize life as we know it, experts are still concerned they could do us more harm than good. Identifying AI’s limitations is one of the best tools we have for seeing how big the gap truly is between machine and human intelligence. > > “Internet memes serve as excellent checkpoints to ensure humans have the upper hand over machines.” On the path to receiving a Ph.D. in electrical and computer engineering at Delaware University in 2021, [Ishaani Priyadarshini](https://www.ischool.berkeley.edu/people/ishaani-priyadarshini) wrote her dissertation on the topic of [AI’s](https://www.popularmechanics.com/technology/security/a45920918/what-are-deepfakes/) inability to decipher memes, and how that could help us in the battle toward singularity. “Internet memes serve as excellent checkpoints to ensure humans have the upper hand over machines, thereby preventing machines from understanding memes and surpassing human intelligence,” she tells *Popular Mechanics*. ## Memes vs. Superintelligent Machines Before pursuing artificial intelligence, Priyadarshini’s background was in [cybersecurity](https://www.popularmechanics.com/technology/security/a40807300/cybersecurity-experts-never-sleep/), and she observed that over 50 percent of the internet is made up of bots—some of these bots are good, while others have [malicious intent](https://www.popularmechanics.com/technology/security/a45533297/what-is-wormgpt/). This confirmed her worst fear: if AI were to achieve sentience and go rogue, it could spell disaster for your online data and privacy. “They would be able to gain every piece of information that is out there in the cyberspace … it could lead to catastrophic results in the future,” says Priyadarshini. Her goal was to find a security measure that could defeat [artificial intelligence](https://www.popularmechanics.com/home/how-to-plans/a44462561/chatgpt-woodworking/) before it ever reaches this point. The solution? Memes. These surprisingly complex amalgamations of text and images often pick apart the more humorous bits of pop culture and current events—and are usually posted across multiple social media platforms. (Picture Grumpy Cat and Doge.) Priyadarshini wanted to find out why AI was so poor at decoding memes, along with how the humorous posts could be used as a cybersecurity tool in the future. AI’s inability to comprehend memes is multifaceted, but much of it has to do with the fact that they tend to be ambiguous. Priyadarshini says that our understanding of memes relies heavily on our own life experiences. “For a machine, memes are merely a bunch of text and images,” she explains. ## 💡 Understand Optical Character Recognition - [The Librarians of the Future Will Be AI Archivists](https://www.popularmechanics.com/technology/a32436235/library-of-congress-machine-learning-newspaper-images/) We already have AIs that are able to recognize text and [facial expressions](https://www.popularmechanics.com/military/a43064899/air-force-drones-facial-recognition/), but there are really no rules with memes. Sure, [Optical Character Recognition](https://www.popularmechanics.com/technology/a32436235/library-of-congress-machine-learning-newspaper-images/) (an advanced system, able to recognize text in images) can already be used to accurately read license plates, but this type of machine learning already has a very good idea of where to look—decoding the text from memes is a completely different story. “There’s no template that memes follow, which is why extracting the text from them is really challenging,” says Priyadarshini. You could argue that most memes are just variations on the same template, but the same meme could still be interpreted wildly differently depending on the content that’s been injected. ## Memes Are Much More Complex Than We Think We currently have no idea if AI has the ability to be sentient, meaning it would be able to think for itself and make its own decisions; even if we did know, it’s unclear how we’d measure just how sentient an AI is. That being said, we *can* measure AI’s ability to extract emotions from text and facial expressions—and it’s already pretty awful at picking up emotions from snippets of text. The billions of permutations you’ll find in both text and facial expressions make it difficult for AI to crack the code. That’s before we even bring in the context of the meme itself. ## 🤖 Feed Your AI Appetite - [Should We Worry About AI Reaching Singularity?](https://www.popularmechanics.com/technology/a42828175/what-is-the-singularity/) - [Deepfakes Are Everywhere—Here’s How to Spot Them](https://www.popularmechanics.com/technology/security/a45920918/what-are-deepfakes/) - [WormGPT: What is it and Should You Be Worried?](https://www.popularmechanics.com/technology/security/a45533297/what-is-wormgpt/) “Understanding memes could forever be a challenge for AI and remain an unsolvable problem,” Priyadarshini says in her [dissertation](https://www.techscience.com/cmc/v70n1/44391/html). Despite memes’ reputation as nothing more than mindless images on the internet, the process of unpacking them is actually much more complex than we thought. The human [brain](https://www.popularmechanics.com/science/a45574179/architecture-of-consciousness/) has an uncanny knack for parsing through an unimaginable amount of data very quickly—helping us understand memes in a matter of seconds. “Superintelligent machines” ... not so much. > ✅ ***Understanding Context*** > Without life experiences, AI doesn’t have much context to lean on. Remember all the > > [Suez Canal](https://www.popularmechanics.com/technology/infrastructure/a36073961/ever-given-ship-stuck-in-suez-canal-timeline/) > > memes from back in 2021? Let’s use the meme below as an example: > [View full post on X](https://twitter.com/ZainaErhaim/status/1374667602308567042) > An AI might be able to recognize that there’s a container ship in the video. However, can it properly identify it as the Ever Given that had run aground in the Suez Canal? To add another layer of complexity on top of that, why is Austin Powers in the background making a 27-point turn? Does it know that the Austin Powers movie series is supposed to be funny? “It may not be an easy task for AI to comprehend internet memes given the plethora and variety of memes found on the internet, along with the different patterns each of the memes can generate,” Priyadarshini says in her dissertation, which was published in the journal > > [Computers, Materials, and Continua](https://www.techscience.com/cmc/v70n1/44391) > > . > Meanwhile, the human brain is able to pick up the pieces with very little effort and register the clever parallel between Austin Powers’ multi-point turn and the container ship maneuvering out of the Suez canal. An [NPR exclusive](https://www.npr.org/2020/07/14/891140598/understanding-unconscious-bias#:~:text=via%20Getty%20Images-,The%20human%20brain%20can%20process%2011%20million%20bits%20of%20information,bits%20of%20information%20a%20second.) with behavioral and data scientist Pragya Agarwal, Ph.D., reveals that the human brain can process roughly 11 million bits of information every second—that’s 0.011 gigabits per second. However, our [conscious mind](https://www.popularmechanics.com/science/a38349973/conscious-brain-multitasking/) (filled with thoughts we’re aware of) can only grab onto about 40 to 50 bits per second. Thankfully, our brain’s amygdala is able to drink from this firehose of information by taking shortcuts to get the bigger picture; Agarwal mentions that the amygdala (the part of your brain responsible for your “gut” feeling) is able to match new information with pre-existing stereotypes and templates to make sense of it. ## AI Databases Are Difficult and Labor Intensive to Train Even if AI could successfully decode memes, that’s only one part of the puzzle. Internet memes are moving at, *well*, the speed of the [internet](https://www.popularmechanics.com/technology/infrastructure/a29627312/how-the-internet-works/)—meaning AI has to keep up. Something that was funny last week could be considered old news in a matter of days. To offer some perspective, [ChatGPT](https://www.popularmechanics.com/technology/a42733497/how-does-chatgpt-work/) (an AI chatbot) has only been trained with a [snapshot of the internet from 2021](https://www.popularmechanics.com/technology/a45359808/gpt4-update/). Anything after that simply doesn’t exist to it. [View full post on X](https://twitter.com/shadesofgreaves/status/1483861298341371907) Think of the memes that we saw in 2022, like those about the distribution of 1 billion free [COVID-19 tests](https://www.popularmechanics.com/science/health/a38816218/free-home-covid-tests/). Sure, the meme involves a photograph of the family from *Full House*, which AI would likely be able to recognize. However, without any context that the U.S. government was giving out free COVID-19 test kits, how would AI know what on Earth was going on in this image? It would have to extrapolate the fact that every home in the United States is only eligible for just four free covid tests. We can clearly deduce that four test kits won’t nearly be enough for most families, but AI isn’t that sophisticated. Plus, artificial intelligence doesn’t really [have a sense of humor.](https://www.popularmechanics.com/technology/robots/a32614583/robot-comedians-better-human-machine-relationships/) ## Humor and AI Don’t Mix (Yet) Being funny is an incredibly sophisticated skill that can mean different things to different people. This subjectivity makes it difficult for AI to understand humor—remember that only certain people might find one piece of content inherently funny. Priyadarshini mentions that AI would be able to parse through screen grabs from *The Office* and know that it’s a comedy, but it wouldn’t be able to unpack the humor within each scene; it has more of a binary *is-this-funny-or-not* type of understanding. ## 💡 Practical Uses for AI - [Why You Should Try Using AI For Your DIY Projects](https://www.popularmechanics.com/home/how-to-plans/a46328877/using-ai-for-diy-projects/) - [I Tried to Learn Woodworking Through ChatGPT](https://www.popularmechanics.com/home/how-to-plans/a44462561/chatgpt-woodworking/) “The real challenge comes from identifying the different kinds of humor that are in those sentences or the videos,” says Priyadarshini. We’re just talking about general humor here, but there are also jokes that might involve dark humor, witty humor, or even dry humor. It really brings into sharp focus the [psychology](https://www.popularmechanics.com/culture/a46211474/liminal-spaces-explained/) of what makes something funny. “We also don’t know what is the true sense of humor … it is different for everyone,” says Priyadarshini. Without AI being sentient, it’s not clear if it will ever be able to have a complete understanding of humor. ## A Futuristic CAPTCHA Riffing on *Full House* is all well and good, but what will this next-level [CAPTCHA](https://www.popularmechanics.com/culture/web/a13163/no-captcha-recatcha-asks-are-you-a-robot-17492127/)—short for Completely Automated Public Turing to tell Computers and Humans Apart—look like in a real-world application? Priyadarshini simply thinks of it as a much more three-dimensional test to differentiate humans from machines; instead of reading weird text or picking which boxes have bridges in them, you’ll instead have to decode the meaning of a meme. She gave a theoretical example of an authentication system for a [nuclear power station](https://www.popularmechanics.com/science/energy/a34976294/tiny-nuclear-reactors/). The employee would begin as usual by entering their username and password. Once that’s done, they’ll be prompted with a meme. In this instance, Priyadarshini used the case of the *Chubby Bubbles* meme (see below) where the user will be given four prompts to identify what is going on in the meme. There are still more questions than answers about how this type of test would work in practice, but it’s a novel idea that could revolutionize the future of cybersecurity. [View full post on X](https://twitter.com/ZatchMehr/status/1571658856111153153) We love the remote possibility of nuclear power station workers deciphering random memes before getting their day going. Sure, we’re most likely not giving up the [nuclear football system](https://www.popularmechanics.com/military/weapons/a35255791/nuclear-football-transfer-trump-biden/) anytime soon, but it certainly poses the question of improving security measures elsewhere. Above all, we’re happy to report the future of cybersecurity is looking bright—and full of hilarious memes.  [Matt Crisara](https://www.popularmechanics.com/author/228667/matt-crisara/) Service Editor Matt Crisara is a native Austinite who has an unbridled passion for cars and motorsports, both foreign and domestic. He was previously a contributing writer for Motor1 following internships at Circuit Of The Americas F1 Track and Speed City, an Austin radio broadcaster focused on the world of motor racing. He earned a bachelor’s degree from the University of Arizona School of Journalism, where he raced mountain bikes with the University Club Team. When he isn’t working, he enjoys sim-racing, FPV drones, and the great outdoors. [Read full bio](https://www.popularmechanics.com/author/228667/matt-crisara/) This content is imported from OpenWeb. You may be able to find the same content in another format, or you may be able to find more information, at their web site. Watch NextThe Unheard Of Roman Sky Camp ![Play video](<data:image/svg+xml,\<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" fill="%23ffffff" viewBox="0 0 256 256"\>%0A%09\<path d="M240,128a15.74,15.74,0,0,1-7.6,13.51L88.32,229.65a16,16,0,0,1-16.2.3A15.86,15.86,0,0,1,64,216.13V39.87a15.86,15.86,0,0,1,8.12-13.82,16,16,0,0,1,16.2.3L232.4,114.49A15.74,15.74,0,0,1,240,128Z"\>%0A\</path\>\</svg\>>)  The Unheard Of Roman Sky Camp Popular Mechanics US ![Playlist Back](<data:image/svg+xml,\<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" fill="%23ffffff" viewBox="0 0 256 256"\>%0A%09\<path d="M208,47.88V208.12a16,16,0,0,1-24.43,13.43L64,146.77V216a8,8,0,0,1-16,0V40a8,8,0,0,1,16,0v69.23L183.57,34.45A15.95,15.95,0,0,1,208,47.88Z"\>\</path\>%0A\</svg\>>) ![Play](<data:image/svg+xml,\<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" fill="%23ffffff" viewBox="0 0 256 256"\>%0A%09\<path d="M240,128a15.74,15.74,0,0,1-7.6,13.51L88.32,229.65a16,16,0,0,1-16.2.3A15.86,15.86,0,0,1,64,216.13V39.87a15.86,15.86,0,0,1,8.12-13.82,16,16,0,0,1,16.2.3L232.4,114.49A15.74,15.74,0,0,1,240,128Z"\>%0A\</path\>\</svg\>>)![Pause](<data:image/svg+xml,\<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" fill="%23ffffff" viewBox="0 0 256 256"\>%0A%09\<path d="M216,48V208a16,16,0,0,1-16,16H160a16,16,0,0,1-16-16V48a16,16,0,0,1,16-16h40A16,16,0,0,1,216,48ZM96,32H56A16,16,0,0,0,40,48V208a16,16,0,0,0,16,16H96a16,16,0,0,0,16-16V48A16,16,0,0,0,96,32Z"\>\</path\>%0A\</svg\>%0A>)![Replay](<data:image/svg+xml,\<svg width="20" height="21" viewBox="0 0 20 21" fill="none" xmlns="http://www.w3.org/2000/svg"\>%0A\<path d="M17.8125 10.5001C17.8127 12.5541 17.004 14.5256 15.5614 15.9877C14.1188 17.4499 12.1585 18.2851 10.1047 18.3126H10C8.00552 18.3168 6.08591 17.5534 4.63906 16.1806C4.45828 16.0099 4.35273 15.7743 4.34562 15.5258C4.33851 15.2772 4.43044 15.036 4.60117 14.8552C4.77191 14.6744 5.00746 14.5689 5.25602 14.5618C5.50458 14.5547 5.74578 14.6466 5.92656 14.8173C6.77519 15.6185 7.84111 16.1518 8.99112 16.3506C10.1411 16.5494 11.3242 16.4049 12.3926 15.9352C13.4609 15.4654 14.3671 14.6912 14.998 13.7093C15.6288 12.7274 15.9563 11.5814 15.9395 10.4144C15.9227 9.24745 15.5623 8.11133 14.9035 7.14802C14.2446 6.1847 13.3165 5.43691 12.2351 4.9981C11.1536 4.55929 9.96684 4.44891 8.82303 4.68076C7.67922 4.91261 6.62911 5.47641 5.80391 6.3017C5.79375 6.31185 5.78438 6.32123 5.77344 6.3306L4.28828 7.68763H5.625C5.87364 7.68763 6.1121 7.78641 6.28791 7.96222C6.46373 8.13804 6.5625 8.37649 6.5625 8.62513C6.5625 8.87377 6.46373 9.11223 6.28791 9.28805C6.1121 9.46386 5.87364 9.56263 5.625 9.56263H1.875C1.62636 9.56263 1.3879 9.46386 1.21209 9.28805C1.03627 9.11223 0.9375 8.87377 0.9375 8.62513V4.87513C0.9375 4.62649 1.03627 4.38804 1.21209 4.21222C1.3879 4.03641 1.62636 3.93763 1.875 3.93763C2.12364 3.93763 2.3621 4.03641 2.53791 4.21222C2.71373 4.38804 2.8125 4.62649 2.8125 4.87513V6.49388L4.49062 4.95795C5.58508 3.86937 6.97748 3.12942 8.49209 2.83148C10.0067 2.53354 11.5756 2.69096 13.0008 3.28388C14.4261 3.87681 15.6437 4.87865 16.5001 6.16296C17.3564 7.44728 17.8131 8.9565 17.8125 10.5001Z" fill="white"/\>%0A\</svg\>%0A>) ![Playlist Forward](<data:image/svg+xml,\<svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" fill="%23ffffff" viewBox="0 0 256 256"\>%0A%09\<path d="M208,40V216a8,8,0,0,1-16,0V146.77L72.43,221.55A15.95,15.95,0,0,1,48,208.12V47.88A15.95,15.95,0,0,1,72.43,34.45L192,109.23V40a8,8,0,0,1,16,0Z"\>\</path\>%0A\</svg\>>) LIVE ![Unmute](<data:image/svg+xml,\<svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" fill="%23ffffff" viewBox="0 0 256 256"\>%0A%09\<path d="M213.92,210.62a8,8,0,1,1-11.84,10.76L160,175.09v48.6a8.29,8.29,0,0,1-3.91,7.18,8,8,0,0,1-9-.56l-65.55-51A4,4,0,0,1,80,176.18V87.09L42.08,45.38A8,8,0,1,1,53.92,34.62Zm-27.21-55.46a8,8,0,0,0,11.29-.7,40,40,0,0,0,0-52.88,8,8,0,1,0-12,10.57,24,24,0,0,1,0,31.72A8,8,0,0,0,186.71,155.16Zm40.92-80.49a8,8,0,1,0-11.92,10.66,64,64,0,0,1,0,85.34,8,8,0,1,0,11.92,10.66,80,80,0,0,0,0-106.66ZM153,119.87a4,4,0,0,0,7-2.7V32.25a8.27,8.27,0,0,0-2.88-6.4,8,8,0,0,0-10-.16L103.83,59.33a4,4,0,0,0-.5,5.85ZM60,80H32A16,16,0,0,0,16,96v64a16,16,0,0,0,16,16H60a4,4,0,0,0,4-4V84A4,4,0,0,0,60,80Z"\>\</path\>%0A\</svg\>>)![Mute](<data:image/svg+xml,\<svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" fill="%23ffffff" viewBox="0 0 256 256"\>%0A%09\<path d="M160,32.25V223.69a8.29,8.29,0,0,1-3.91,7.18,8,8,0,0,1-9-.56l-65.57-51A4,4,0,0,1,80,176.16V79.84a4,4,0,0,1,1.55-3.15l65.57-51a8,8,0,0,1,10,.16A8.27,8.27,0,0,1,160,32.25ZM60,80H32A16,16,0,0,0,16,96v64a16,16,0,0,0,16,16H60a4,4,0,0,0,4-4V84A4,4,0,0,0,60,80Zm126.77,20.84a8,8,0,0,0-.72,11.3,24,24,0,0,1,0,31.72,8,8,0,1,0,12,10.58,40,40,0,0,0,0-52.88A8,8,0,0,0,186.74,100.84Zm40.89-26.17a8,8,0,1,0-11.92,10.66,64,64,0,0,1,0,85.34,8,8,0,1,0,11.92,10.66,80,80,0,0,0,0-106.66Z"\>\</path\>%0A\</svg\>>) Current Time 0:00 Total Duration 0:51 ![Closed Captions](<data:image/svg+xml,\<svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" fill="%23ffffff" viewBox="0 0 256 256"\>%0A%09\<path d="M224,48H32A16,16,0,0,0,16,64V192a16,16,0,0,0,16,16H224a16,16,0,0,0,16-16V64A16,16,0,0,0,224,48ZM116,162.64a40,40,0,1,1,0-69.28,8,8,0,1,1-8,13.85,24,24,0,1,0,0,41.58,8,8,0,0,1,8,13.85Zm80,0a40,40,0,1,1,0-69.28,8,8,0,1,1-8,13.85,24,24,0,1,0,0,41.58,8,8,0,0,1,8,13.85Z"\>\</path\>%0A\</svg\>>) - Captions Off - English ![Enter Fullscreen](<data:image/svg+xml,\<svg width="20" height="21" viewBox="0 0 20 21" fill="none" xmlns="http://www.w3.org/2000/svg"\>%0A\<path d="M4.0415 7.79134H1.6665V1.83301H7.62484V4.20801H4.0415V7.79134ZM1.6665 12.5413H4.0415V16.1247H7.62484V18.4997H1.6665V12.5413ZM15.9582 4.20801H12.3748V1.83301H18.3332V7.79134H15.9582V4.20801ZM12.3748 18.4997V16.1247H15.9582V12.5413H18.3332V18.4997H12.3748Z" fill="white"/\>%0A\</svg\>%0A>)![Exit Fullscreen](<data:image/svg+xml,\<svg width="20" height="21" viewBox="0 0 20 21" fill="none" xmlns="http://www.w3.org/2000/svg"\>%0A\<path d="M1.6665 5.73794H5.23794V2.1665H7.6189V8.1189H1.6665V5.73794ZM5.23794 15.2617H1.6665V12.8808H7.6189V18.8332H5.23794V15.2617ZM12.3808 2.1665H14.7617V5.73794H18.3332V8.1189H12.3808V2.1665ZM14.7617 15.2617V18.8332H12.3808V12.8808H18.3332V15.2617H14.7617Z" fill="white"/\>%0A\</svg\>%0A>) Watch: [The Unheard Of Roman Sky Camp](https://www.popularmechanics.com/science/archaeology/a70744334/roman-army-camp-alps-hidden-in-sky/) ✕ Advertisement - Continue Reading Below ## [Artificial Intelligence](https://www.popularmechanics.com/artificial-intelligence/) [This Habit Could Warp Your Sense of Reality](https://www.popularmechanics.com/science/a69928076/cognitive-corridor-ai-learning-atrophy/) [We Put Photoshop’s Generative AI to the Test](https://www.popularmechanics.com/technology/apps/a44764303/ai-photo-editor-photoshop-generative-fill/) [A Fully Autonomous Ship's Perilous Journey at Sea](https://www.popularmechanics.com/technology/infrastructure/a44936536/autonomous-mayflower/) [Elon Musk Launches His Own AI Company: xAI](https://www.popularmechanics.com/technology/robots/a44535534/elon-musk-xai/) Advertisement - Continue Reading Below [Why AI May Be at the Heart of the Reddit Protest](https://www.popularmechanics.com/technology/apps/a44202232/reddit-protest-explained/) [The Popular Mechanics Guide to AI Music Generators](https://www.popularmechanics.com/technology/audio/a44109081/ai-music-generators-explained/) [Experts Warn that AI Could Cause Human Extinction](https://www.popularmechanics.com/technology/robots/a44041542/ai-extinction/) [Microsoft Says GPT-4 Shows Signs of Human Reason](https://www.popularmechanics.com/technology/robots/a43906996/artificial-intelligence-shows-signs-of-human-reasoning/) [Russia's Felon Is Getting an AI Boost](https://www.popularmechanics.com/military/aviation/a43689560/russia-su-57-felon-fighter-jet-artificial-intelligence/) [Top 3 AI Image Generators](https://www.popularmechanics.com/technology/a43557128/ai-image-generators-review/) [Scientist: Superhuman AI May Eliminate Humanity](https://www.popularmechanics.com/technology/a43441905/superhuman-ai-may-eliminate-humanity/) [A Chernobyl for AI May Be on the Horizon](https://www.popularmechanics.com/technology/a43540293/chernobyl-for-artificial-intelligence/) Advertisement - Continue Reading Below [](https://www.popularmechanics.com/) [](https://x.com/PopMech) [](https://www.tiktok.com/@popularmechanics?lang=en) [](https://www.youtube.com/c/popularmechanics?sub_confirmation=1) [](https://www.facebook.com/popularmechanics) [](https://www.instagram.com/popularmechanics/) [](https://www.pinterest.com/popmech/?auto_follow=1) [Newsletter](https://www.popularmechanics.com/newsletter)[About Us](https://www.popularmechanics.com/about/a45/about-us/)[Media Kit](http://www.popularmechanicsmediakit.com/r5/home.asp)[Writer Guidelines](https://www.popularmechanics.com/about/a34246962/how-to-pitch-popular-mechanics/)[Press Room](http://www.hearst.com/news)[Contact Us](https://www.popularmechanics.com/about/a2922/contact-us/)[Community Guidelines](https://www.popularmechanics.com/about/a3024/community-guidelines/)[Advertise Online](https://www.advertising.hearstmagazines.com/brands/popular-mechanics)[Customer Service](https://www.popularmechanics.com/service)[Subscribe](https://shop.popularmechanics.com/popular-mechanics-all-access-3.html?cds_tracking_code=popfooter)[Digital Issues](https://www.popularmechanics.com/promotions/a40118846/download-exclusive-pop-mech-digital-issues/)[PopMech Pro Membership Perks](https://www.popularmechanics.com/about/a46147319/pop-mech-pro-membership/)  A Part of Hearst Digital Media We may earn commission from links on this page, but we only recommend products we back. ©2026 Hearst Magazine Media, Inc. All Rights Reserved. [Privacy Notice](https://www.hearst.com/-/us-magazines-privacy-notice)[CA Notice at Collection](https://www.hearst.com/-/us-magazines-privacy-notice#_ADDITIONAL_INFO)[Your CA Privacy Rights/Shine the Light](https://www.hearst.com/-/us-magazines-privacy-notice#_ADDITIONAL_INFO)[DAA Industry Opt Out](https://www.hearst.com/-/us-magazines-privacy-notice#_OPT_OUTS)[Terms of Use](https://www.hearst.com/-/us-magazines-terms-of-use)[Site Map](https://www.popularmechanics.com/sitemap/) [Cookies Choices](https://www.popularmechanics.com/technology/security/a43013341/why-ai-doesnt-understand-memes/ "Cookies Choices")

Artificial intelligence is supposedly drawing closer to [singularity](https://www.popularmechanics.com/technology/a42828175/what-is-the-singularity/)—the point at which machine intelligence surpasses human intelligence. That’s sparked debates about how we can best control AI’s capabilities; while superintelligent systems *could* revolutionize life as we know it, experts are still concerned they could do us more harm than good. Identifying AI’s limitations is one of the best tools we have for seeing how big the gap truly is between machine and human intelligence. > > “Internet memes serve as excellent checkpoints to ensure humans have the upper hand over machines.” On the path to receiving a Ph.D. in electrical and computer engineering at Delaware University in 2021, [Ishaani Priyadarshini](https://www.ischool.berkeley.edu/people/ishaani-priyadarshini) wrote her dissertation on the topic of [AI’s](https://www.popularmechanics.com/technology/security/a45920918/what-are-deepfakes/) inability to decipher memes, and how that could help us in the battle toward singularity. “Internet memes serve as excellent checkpoints to ensure humans have the upper hand over machines, thereby preventing machines from understanding memes and surpassing human intelligence,” she tells *Popular Mechanics*. ## Memes vs. Superintelligent Machines Before pursuing artificial intelligence, Priyadarshini’s background was in [cybersecurity](https://www.popularmechanics.com/technology/security/a40807300/cybersecurity-experts-never-sleep/), and she observed that over 50 percent of the internet is made up of bots—some of these bots are good, while others have [malicious intent](https://www.popularmechanics.com/technology/security/a45533297/what-is-wormgpt/). This confirmed her worst fear: if AI were to achieve sentience and go rogue, it could spell disaster for your online data and privacy. “They would be able to gain every piece of information that is out there in the cyberspace … it could lead to catastrophic results in the future,” says Priyadarshini. Her goal was to find a security measure that could defeat [artificial intelligence](https://www.popularmechanics.com/home/how-to-plans/a44462561/chatgpt-woodworking/) before it ever reaches this point. The solution? Memes. These surprisingly complex amalgamations of text and images often pick apart the more humorous bits of pop culture and current events—and are usually posted across multiple social media platforms. (Picture Grumpy Cat and Doge.) Priyadarshini wanted to find out why AI was so poor at decoding memes, along with how the humorous posts could be used as a cybersecurity tool in the future. AI’s inability to comprehend memes is multifaceted, but much of it has to do with the fact that they tend to be ambiguous. Priyadarshini says that our understanding of memes relies heavily on our own life experiences. “For a machine, memes are merely a bunch of text and images,” she explains. We already have AIs that are able to recognize text and [facial expressions](https://www.popularmechanics.com/military/a43064899/air-force-drones-facial-recognition/), but there are really no rules with memes. Sure, [Optical Character Recognition](https://www.popularmechanics.com/technology/a32436235/library-of-congress-machine-learning-newspaper-images/) (an advanced system, able to recognize text in images) can already be used to accurately read license plates, but this type of machine learning already has a very good idea of where to look—decoding the text from memes is a completely different story. “There’s no template that memes follow, which is why extracting the text from them is really challenging,” says Priyadarshini. You could argue that most memes are just variations on the same template, but the same meme could still be interpreted wildly differently depending on the content that’s been injected. ## Memes Are Much More Complex Than We Think We currently have no idea if AI has the ability to be sentient, meaning it would be able to think for itself and make its own decisions; even if we did know, it’s unclear how we’d measure just how sentient an AI is. That being said, we *can* measure AI’s ability to extract emotions from text and facial expressions—and it’s already pretty awful at picking up emotions from snippets of text. The billions of permutations you’ll find in both text and facial expressions make it difficult for AI to crack the code. That’s before we even bring in the context of the meme itself. “Understanding memes could forever be a challenge for AI and remain an unsolvable problem,” Priyadarshini says in her [dissertation](https://www.techscience.com/cmc/v70n1/44391/html). Despite memes’ reputation as nothing more than mindless images on the internet, the process of unpacking them is actually much more complex than we thought. The human [brain](https://www.popularmechanics.com/science/a45574179/architecture-of-consciousness/) has an uncanny knack for parsing through an unimaginable amount of data very quickly—helping us understand memes in a matter of seconds. “Superintelligent machines” ... not so much. > ✅ ***Understanding Context*** > Without life experiences, AI doesn’t have much context to lean on. Remember all the > > [Suez Canal](https://www.popularmechanics.com/technology/infrastructure/a36073961/ever-given-ship-stuck-in-suez-canal-timeline/) > > memes from back in 2021? Let’s use the meme below as an example: > An AI might be able to recognize that there’s a container ship in the video. However, can it properly identify it as the Ever Given that had run aground in the Suez Canal? To add another layer of complexity on top of that, why is Austin Powers in the background making a 27-point turn? Does it know that the Austin Powers movie series is supposed to be funny? “It may not be an easy task for AI to comprehend internet memes given the plethora and variety of memes found on the internet, along with the different patterns each of the memes can generate,” Priyadarshini says in her dissertation, which was published in the journal > > [Computers, Materials, and Continua](https://www.techscience.com/cmc/v70n1/44391) > > . > Meanwhile, the human brain is able to pick up the pieces with very little effort and register the clever parallel between Austin Powers’ multi-point turn and the container ship maneuvering out of the Suez canal. An [NPR exclusive](https://www.npr.org/2020/07/14/891140598/understanding-unconscious-bias#:~:text=via%20Getty%20Images-,The%20human%20brain%20can%20process%2011%20million%20bits%20of%20information,bits%20of%20information%20a%20second.) with behavioral and data scientist Pragya Agarwal, Ph.D., reveals that the human brain can process roughly 11 million bits of information every second—that’s 0.011 gigabits per second. However, our [conscious mind](https://www.popularmechanics.com/science/a38349973/conscious-brain-multitasking/) (filled with thoughts we’re aware of) can only grab onto about 40 to 50 bits per second. Thankfully, our brain’s amygdala is able to drink from this firehose of information by taking shortcuts to get the bigger picture; Agarwal mentions that the amygdala (the part of your brain responsible for your “gut” feeling) is able to match new information with pre-existing stereotypes and templates to make sense of it. ## AI Databases Are Difficult and Labor Intensive to Train Even if AI could successfully decode memes, that’s only one part of the puzzle. Internet memes are moving at, *well*, the speed of the [internet](https://www.popularmechanics.com/technology/infrastructure/a29627312/how-the-internet-works/)—meaning AI has to keep up. Something that was funny last week could be considered old news in a matter of days. To offer some perspective, [ChatGPT](https://www.popularmechanics.com/technology/a42733497/how-does-chatgpt-work/) (an AI chatbot) has only been trained with a [snapshot of the internet from 2021](https://www.popularmechanics.com/technology/a45359808/gpt4-update/). Anything after that simply doesn’t exist to it. Think of the memes that we saw in 2022, like those about the distribution of 1 billion free [COVID-19 tests](https://www.popularmechanics.com/science/health/a38816218/free-home-covid-tests/). Sure, the meme involves a photograph of the family from *Full House*, which AI would likely be able to recognize. However, without any context that the U.S. government was giving out free COVID-19 test kits, how would AI know what on Earth was going on in this image? It would have to extrapolate the fact that every home in the United States is only eligible for just four free covid tests. We can clearly deduce that four test kits won’t nearly be enough for most families, but AI isn’t that sophisticated. Plus, artificial intelligence doesn’t really [have a sense of humor.](https://www.popularmechanics.com/technology/robots/a32614583/robot-comedians-better-human-machine-relationships/) ## Humor and AI Don’t Mix (Yet) Being funny is an incredibly sophisticated skill that can mean different things to different people. This subjectivity makes it difficult for AI to understand humor—remember that only certain people might find one piece of content inherently funny. Priyadarshini mentions that AI would be able to parse through screen grabs from *The Office* and know that it’s a comedy, but it wouldn’t be able to unpack the humor within each scene; it has more of a binary *is-this-funny-or-not* type of understanding. “The real challenge comes from identifying the different kinds of humor that are in those sentences or the videos,” says Priyadarshini. We’re just talking about general humor here, but there are also jokes that might involve dark humor, witty humor, or even dry humor. It really brings into sharp focus the [psychology](https://www.popularmechanics.com/culture/a46211474/liminal-spaces-explained/) of what makes something funny. “We also don’t know what is the true sense of humor … it is different for everyone,” says Priyadarshini. Without AI being sentient, it’s not clear if it will ever be able to have a complete understanding of humor. ## A Futuristic CAPTCHA Riffing on *Full House* is all well and good, but what will this next-level [CAPTCHA](https://www.popularmechanics.com/culture/web/a13163/no-captcha-recatcha-asks-are-you-a-robot-17492127/)—short for Completely Automated Public Turing to tell Computers and Humans Apart—look like in a real-world application? Priyadarshini simply thinks of it as a much more three-dimensional test to differentiate humans from machines; instead of reading weird text or picking which boxes have bridges in them, you’ll instead have to decode the meaning of a meme. She gave a theoretical example of an authentication system for a [nuclear power station](https://www.popularmechanics.com/science/energy/a34976294/tiny-nuclear-reactors/). The employee would begin as usual by entering their username and password. Once that’s done, they’ll be prompted with a meme. In this instance, Priyadarshini used the case of the *Chubby Bubbles* meme (see below) where the user will be given four prompts to identify what is going on in the meme. There are still more questions than answers about how this type of test would work in practice, but it’s a novel idea that could revolutionize the future of cybersecurity. We love the remote possibility of nuclear power station workers deciphering random memes before getting their day going. Sure, we’re most likely not giving up the [nuclear football system](https://www.popularmechanics.com/military/weapons/a35255791/nuclear-football-transfer-trump-biden/) anytime soon, but it certainly poses the question of improving security measures elsewhere. Above all, we’re happy to report the future of cybersecurity is looking bright—and full of hilarious memes.  Matt Crisara is a native Austinite who has an unbridled passion for cars and motorsports, both foreign and domestic. He was previously a contributing writer for Motor1 following internships at Circuit Of The Americas F1 Track and Speed City, an Austin radio broadcaster focused on the world of motor racing. He earned a bachelor’s degree from the University of Arizona School of Journalism, where he raced mountain bikes with the University Club Team. When he isn’t working, he enjoys sim-racing, FPV drones, and the great outdoors. [Read full bio](https://www.popularmechanics.com/author/228667/matt-crisara/) This content is imported from OpenWeb. You may be able to find the same content in another format, or you may be able to find more information, at their web site.