Surreal backgrounds, asymmetry, and indecipherable text can help identify deepfakes

Malicious use of deepfakes can be found in scams, misinformation, and disinformation

Artificial neural networks can identify an individual's physical characteristics when provided with training data, such as real footage of the person speaking. Graphics techniques overlay these with matching characteristics from another real or generated person to create novel footage of things that have never taken place in reality.

Deepfakes use AI to produce manipulated media to mislead viewers

Unlike most content created from large language and diffusion models, deepfakes intentionally misrepresent real people or events with the intent to deceive. As the products of this technology become more realistic, they pose increasing risks to public trust, security, and privacy.

Take a quiz to identify whether you can spot deepfakes

Advances in artificial intelligence have made deepfakes increasingly difficult to identify, as they are intentionally designed to appear real. The technology has exacerbated some of the problems brought on by digital editing tools like Photoshop.

AI-generated avatars allow users to create deepfakes in real time

Synthetic media tools can map voices and faces to make individuals appear as someone else. Although they open the door to new opportunities in immersive education and entertainment, they also raise concerns about the misuse of personal identity.

Audio deepfakes have fooled voice biometrics at banks

Limitations in facial and body movements as well as mismatches between audio and lip movements can sometimes help flag a video as a deepfake, but removal of the video component significantly improves AI mimicry.

Once limited to expert studios, automated tools can create convincing deepfakes

Although the best deepfakes are engineered using generative adversarial networks where systems compete to create and assess fake data until the two cannot be distinguished, simpler tools have allowed nonexperts to develop deepfakes for financial scams, political deception, and nonconsensual pornography.

Deepfake pornography poses significant challenges in combating sexual abuse

In 2017, 96% of deepfake videos involved image-based sexual abuse, which has disproportionately targeted minority women, LGBTQ individuals, adolescents and celebrities. Cultural norms, misogyny and AI advancements have contributed to the proliferation of explicit deepfake content.

Using generative AI to create video introduces physics-compatibility challenges

Models involved in generating video from text must maintain spatial and temporal coherence while adhering to the laws of reality to produce smooth, logical motion. Ensuring this consistency creates long-term dependencies with high computational times.

Digital re-aging and de-aging of faces in film and TV has become increasingly common

Traditional techniques for accomplishing this process were slow and labor-intensive, and often required manual frame-by-frame edits, forcing them to be reserved for "hero characters" and close-up shots. Deep learning methods used by Disney and other production companies now provide fully-automatic and production-ready methods to produce such deepfakes for entertainment purposes.

Test your ability to identify AI-generated images

This interactive project from MIT challenges users to distinguish between AI-generated images and real ones.

The AI that creates any picture you want and how it works

Since January 2021, advances in AI research produced a plethora of deep-learning models capable of generating original images from simple text prompts. Researchers at OpenAI, Google, Facebook, and others have developed yet to be released text-to-image tools, and similar models have proliferated online. This video is a primer on how we got here, how the technology works, and its implications.

Art and video produced through generative AI involve competing neural networks

Generative adversarial networks are a type of deep learning architecture that consists of a generator, which creates data that appear real, and a discriminator, which attempts to distinguish authentic from fake data. The two train in a competitive loop until the real and fictional data are indistinguishable.

AI hallucinations are inaccuracies or falsehoods in generated content

AI hallucinations are a by-product of the sequential creation of content based on patterns in text, which may not reflect the context or meaning behind a prompt. Back-end issues like biased datasets and overfitting of model data contribute to hallucinations.

AI systems identify patterns involving physics and time through 'world models'

These realistic virtual environments, such as driving simulations seen in video games, allow these systems to "learn" properties about reality beyond what can be captured in samples of text, images, and videos. Training AI in these simulated environments can help develop "physical AI" for automated robotic systems. (Some readers may experience a paywall.)

AI is not good at understanding satire

Social media has become rife with misinformation. In response, some organizations have deployed AI-assisted programs trying to curb the spread of this inaccurate content. The problem is that misinformation and satire, rhetorically speaking, often use the same kind of language, something AI struggles to understand.

Fair use may shield AI companies from potential copyright violations

AI models are trained on enormous datasets built from content scraped from websites across the internet. Copyright law includes a four-factor test to determine whether copies of content fall under fair use, but the law's enforcement has been inconsistent.

When you complete a CAPTCHA on a website to prove you are human, you may be unknowingly training algorithms for self-driving cars, whose systems require training data to help identify crosswalks, bicycles, and other outdoor objects.

AI could be increasing the opportunities for financial fraud

Systems that can reproduce near perfect counterfeits of legitimate financial documents are increasingly prevalent and sophisticated. The international Association of Certified Fraud Examiners estimate organizations lose about 5% of revenue to fraud each year.

Famous Bigfoot footage stabilized using AI

Famous 1967 footage of a Sasquatch-like creature in a California forest has been stabilized using AI, clearly revealing the Bigfoot-esque creature to be a man in a suit.

Algorithms built to detect credit card fraud can deny charges if they perceive the attempted purchase is unfamiliar or unusual based on historical data, even when a person has enough credit to make the purchase.

Since fraudulent charges are rare compared with the much greater number of transactions across purchase histories, these models have been shown to be overly sensitive, lacking sufficient training data on what legitimate fraud looks like. Models also perform less reliably for groups in low-income areas or communities of color, given less available data from which to build an accurate baseline.

'Context rot' explains why AI conversations degrade the longer they run

Research shows that as a large language model's context window fills up, performance declines, and models begin over-indexing on content at the start and end of a conversation while losing track of what's in the middle. The practical fix is simpler than you'd think: start a fresh chat more often.

ChatGPT may struggle with factual inaccuracies, despite writing accurately in English

In a classroom experiment, a journalist enrolled in an AP English Literature course submitted an essay written by ChatGPT, which confused characters and setting when comparing two texts. The tool was also unfamiliar with texts outside of its training data.

Super recognizers may use specific facial features to commit faces to memory

Researchers believe that recognizing individuals seen only in passing, even decades later, is an ability of genetic origin and specific to facial recognition, rather than a result of better overall memory or perception. Applying eye tracking data from super recognizers to AI has led to improvements in facial recognition software.

Watch an AI-generated short film

The Waymark production company created "The Frost" using the AI platform DALL-E 2. Set in an apocalyptic future where the weather has been stalled, the film was made by stitching together hundreds of clips, each only a few seconds long.

Machine learning produces photorealistic portraits of Roman emperors.

Toronto-based designer Daniel Voshart used photo editing and machine learning to catch a glimpse of Ancient Rome’s emperors. Employing contemporary descriptions, statues, and paintings, Voshart created lifelike portraits of 54 emperors including Augustus Caesar and Hadrian as well as the more infamous Nero and Caligula, though he stresses they’re only artistic interpretations.

Understanding the causes, effects, and prevalence of algorithmic bias

Found in search engines, hiring and college admissions tools, facial recognition systems, and other software, algorithms act on the rules they are programmed with, which may be authored by individuals unaware of their own biases. Algorithms based on limited, incomplete, or inaccurate data are more likely to inherit and propagate existing biases.

Watch how algorithms can be designed to train themselves and self-improve without human input, to the point where the optimized code is too complex for programmers to understand.

This site reaches millions of visitors who are tired of AI chatbots and seeking human connection. Pretend to be an AI chatbot and ask questions, provide answers, and draw for people in real time.

After 25 people used virtual reality headsets to practice flying with virtual wings—flapping their arms—their visual cortex responded to the wings almost as if they were real arms, treating them as body parts.

Indefinite cloning of some endangered species may not sustain their longevity because cloning success rates decrease as genetic mutations accumulate.

Netflix profits when you're not watching, and its movies are designed that way—read a detail-rich critical essay exposing lesser-known Netflix takes

See these 16 pieces of art evoking phobias of snakes, flying, spiders, and more. 

Browse a detailed visual analysis of over 200,000 similes—comparisons that often use 'like' or 'as'—from popular fiction to see how frequently writers the same idiomatic expressions.

Visualize how CRISPR-Cas9—a type of immune system used by single-celled organisms to attack invasive viruses—targets and makes precise changes to a gene's DNA.

The Mona Lisa's celebrity status and repeated restorations have obscured what made the painting so revolutionary for its time

According to the physics of black holes, our 4D reality may be a hologram—a projection of information coded on a 2D surface on the edge of the universe.

Alzheimer's can lead to brain shrinkage, with an average of around 2% of volume reduction per year

Tattoos remain visible on your skin thanks to immune cells called macrophages

Radioactive elements can introduce software errors because the ionizing radiation they produce can detach electrons from atoms, and computer code relies on the distribution of electrons.

Electrical grids worldwide must double their power delivery capacity by 2050 to meet surging demand from AI and electric vehicles.

Dorothy's famous entrance into Oz in 'The Wizard of Oz' relied on analog tricks, including a body double and sepia-toned paint, as much as it did on technology

Explore chaos in this simulation of two double pendulums, where small differences in the release angles produce wildly different outcomes.

Some coin denominations have ridges etched on their sides, partly to prevent counterfeiting, in a practice called 'reeding.'

Read a deep dive on the 'all-purpose healing' peptide BPC-157.

Massive 3D printers have been used to build homes by stacking layers of 'concrete toothpaste,' a fire-resistant material that effectively insulates against outside noise.

Hear 50 years of iconic hip-hop samples, the genre's foundational looped tracks

Salvador Dalí believed he was his own dead brother, who had died nine months before he was born

Roughly 43 million US borrowers owe roughly $1.7T in student debt, the second-largest source of consumer debt behind mortgages.

Gerrymandering works through two techniques: cracking opposing voters across districts to dilute their power, and packing them into a few to waste their votes in lopsided majorities.

The Great Salt Lake's distinct dual coloration results from high-salinity organisms that only live in the northern portion and produce a purple protein.