Scientists increasingly can't predict what AI will do

Decoding the hype around AI

2022 was the year AI broke onto the scene, with applications like ChatGPT amazing users with its ability to generate realistic-sounding answers. Technologist Arvind Narayanan presents a clear-eyed view of what the technologies can (and can't) do.

Google invented the foundation of the AI movement—but ChatGPT was a shock

For years, Google was the cutting-edge of the AI world, including the development of the Transformer—the basis of contemporary AI chatbots. But since OpenAI debuted ChatGPT in 2022, the company was had to play catch-up. This is a four-hour deep dive into Google's AI history.

Anthropic's interpretability team discusses how AI models' thinking mimics biology

Just as humans engage in complex behaviors to survive and reproduce, teams of experts in neuroscience, virology, mathematics, and other disciplines argue that large language models develop complex mechanisms to achieve their goals. These mechanisms can be manipulated to assess how they modify outputs—much like stimulating individual neurons—to uncover how LLMs "think."

Sam Altman says OpenAI released ChatGPT early as a public service

As the company was developing ChatGPT-4, version 3 was released to allow everyone to adjust to the new tool. Since then, he and his colleagues have called for oversight and regulation to address fears of misuse and AI-enabled pathogens.

Research suggests using AI for schoolwork may harm critical thinking skills

An MIT study finds lower brain engagement in students who used ChatGPT, resulting in consistent underperformance in neural, linguistic and behavioral levels. AI users struggled to recall their own essays, while the brain-only group showed the highest neural connectivity and satisfaction.

ChatGPT is trained on most of the world's digitally accessible text

As with other AI chatbots, ChatGPT is primarily trained through non-supervised pretraining, which identifies underlying structure and patterns within input data. Then, human intervention addresses harmful or biased patterns that should not be propagated to generated outputs.

Generative AI tools excel at pattern recognition, not contextual accuracy

Large language models are trained on vast amounts of unstructured data, from which they develop parameters for grammar and associations between words. These connections can introduce errors due to inapplicable reasoning when used on new data in unfamiliar contexts.

Agentic AI systems can proactively achieve user goals without user direction

Unlike large language models, which do not take follow-up actions after generating their output, agentic AI uses an initial prompt to take multiple steps, learn from its environment and outcomes, and continue working without follow-up human prompting.

As of 2025, artificial intelligence systems remain narrow and specialized

Efforts are ongoing to create artificial general intelligence, where machines outperform humans across a wide range of cognitive tasks, including those for which they have not been specifically trained. Current models utilize machine learning and neural networks to perform tasks for which they are programmed, leveraging provided training data.

What does it mean for AI to "understand" something?

This Quanta article discusses how programmers are working to create a reliable test for the depth and extent of the knowledge of machines. What does AI know and how does it know it, if it knows anything at all? There's been progress, but it hasn't been easy. Even humans can still be fooled some of the time.

Google’s CEO has warned that today’s AI boom shows ‘elements of irrationality’

An interview with Alphabet CEO Sundar Pichai warns that the trillion-dollar surge in AI valuations may contain clear signs of irrationality, echoing past tech bubbles. He says no firm—including Google—would be immune if the boom cools or bursts.

Indirect prompt injection is a major security flaw of generative AI systems

These attacks can manipulate AI behavior by hiding instructions in websites and PDFs that become part of training data. The more companies connect sites, services, and sensitive datasets to AI tools, the greater the chances of exposure to malicious code.

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.

Unlike the models made by large AI companies, open-weight models can have their safeguards removed or weakened, a process that has become increasingly easy and may expose users to harmful information.

This removal is made possible by developers making the parameters behind their models publicly known. By adjusting these parameters, users can modify how the model processes and provides information. Many of these models can be downloaded locally, preventing developers from seeing how people are using them.

AI tools are trained on human-created data containing implicit and explicit biases

Because AI tools reflect patterns in their training data, their outputs may reinforce existing societal biases—especially racial ones—rather than offering neutral or fair ones. Adequate training is necessary to undermine these biases whenever new data is incorporated.

AI excels at assisting with structured tasks involving logistics and planning

Tools like ChatGPT, Gemini, and Perplexity can help organize chaotic calendars, troubleshoot bedtime routines, and assist with salary planning and moving. However, specific quotes, like insurance, and local cost comparisons can be flawed since AI draws from nonlocalized datasets.

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.

The do's and don'ts of using AI to manage your finances

Experts suggest avoiding asking AI for investment advice, but say it's useful for basic budgeting help. Read this article to learn more.

AI models incorporate a randomness parameter when generating responses to prompts

AI models use this parameter to prevent repeated outputs by sometimes choosing less likely next words during sequential generation. However, this randomness—alongside insufficient data and training—may cause hallucinations of incorrect results.

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.

Experts suggest there is a 10% chance AI will cause human extinction

AI superintelligence may hijack infrastructure, including power grids and transportation systems, if it surpasses human controls. It may also design unstoppable pathogens or deplete resources, like electricity, meant for human use.

AI data center growth has created shortages and price hikes in computer parts

The parallel processing needed to power large language models for AI chatbots and other tools drove demand for high-speed graphics processing units to spike in 2025, enabling capital-rich technology companies to price out consumers. Similar demands for the computer memory used in storage drives are expected to keep shortages from easing through 2028.

AI griefbots are controversial and could conceivably change how we mourn

"Deathbots" or "griefbots" are a particularly controversial use of AI chatbots. Mourners chat with AI chatbots that generate text or even video avatars of the dead based on inputs from loved ones to theoretically provide closure and comfort.

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.

How to prevent a worst-case scenario with synbio

Author and entrepreneur Rob Reid reviews the risks of a world where more and more people have access to the technology needed to create a doomsday bug that could wipe out humanity.

There’s an artificial intelligence model trained on bankruptcy law

Although it's designed for lawyers, the tool is able to help everyday browsers synthesize important cases and point users to relevant documents about them. One sample using it asks Chatlaw to explain the significance of a recent case and then to define relevant terminology.

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.

A vending machine run with Anthropic's AI went out of business after it was convinced to give away items for free and sell a video game console.

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.

English mathematician Ada Lovelace wrote the first computer algorithm a century before computers existed and hypothesized that computing devices would one day be able to compose music—foreseeing the generative AI systems of today.

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

Chaos theory—informally known as the butterfly effect—is the reason why weather forecasts are presented with probabilities rather than certainties.

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.

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.

In 2025, four executives from large tech companies were given positions in the US Army Reserves—including Palantir Chief Technology Officer Shyam Sankar, which reinforced Palantir's ties with the US military and its role as a leading defense partner.

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

Recent research found that the brain doesn't seem to massively rewire following amputation—upending earlier beliefs about the cause of phantom limb pain.

Clever Hans was a 1900s horse that was believed to be able to solve math equations, tell time, and read and comprehend German, until a psychology student discovered how he got his answers.

After NASA's Voyager 1 captured Earth as a single pixel as it left the solar system in 1990, astronomer Carl Sagan used the image to reflect on humanity's fragility and the responsibility to protect 'the only world we've ever known.'

As of 2025, the most cited papers in science include a review of graphene, methods to improve image recognition, streamlining electron density calculations, and optimizing protein studies.

In recent years a child anxiety expert has created and tested a novel phobia and anxiety treatment approach for children, focused on changing how parents respond to the fears.

Listen to an in-depth conversation on the links between Parkinson's disease and herbicide exposure and the challenging pace of scientific investigation.

The animatronic shark from Steven Spielberg's 'Jaws,' the first modern blockbuster, was a complex underwater apparatus called 'Bruce.'

A Yale student wrote a profound essay about the consulting industry's influence on elite college students that was published in a book of her writing after she died in a post-graduation car accident.

Apeirophobia is an overwhelming fear of living forever, perhaps stemming from the brain's inability to conceptualize an infinite life.

Learn how to read a prospectus, the juicy document that any company must file with the SEC before going public.

Art theft is the third-largest source of criminal revenues worldwide, thanks in part to the subjective nature of valuing a work of art.

Famed poet TS Eliot rejected the manuscript for George Orwell's political fable 'Animal Farm,' writing that its criticism of authoritarianism was not 'the right point of view.'

Most of the observable universe's volume consists of voids and supervoids—enormous regions of space with little to no galaxies, stars, or light.