Computer Science News

Lu et al. suggest large language models like ChatGPT cannot learn independently or acquire new skills.

Large Language Models Don’t Just Analyze People, They Judge Them

Apr 13, 2026 by News Staff

New research from the Hebrew University of Jerusalem shows that large language models (LLMs) form structured ‘trust’ assessments much like humans do,...

Google’s quantum computer. Image credit: Google.

Breakthrough in Quantum Computing: First Secure Method to Back Up Quantum Information

Jan 7, 2026 by News Staff

Scientists at the University of Waterloo and Kyushu University have developed the first method to create redundant, encrypted copies of qubits —...

This conceptual illustration of a computer based on 2D molecules displays an actual scanning electron microscope image of the computer fabricated by Ghosh et al. Image credit: Krishnendu Mukhopadhyay / Penn State.

Researchers Develop World’s First Non-Silicon 2D Computer

Jun 18, 2025 by News Staff

Silicon has enabled advancements in semiconductor technology through miniaturization, but scaling challenges necessitate the exploration of new materials....

A rendering of the ion trap chip. Image credit: Jochen Wolf & Tom Harty, University of Oxford.

Scientists Set New Record for Qubit Operation Accuracy

Jun 10, 2025 by News Staff

A team of physicists at the University of Oxford has achieved the lowest-ever error rate for a quantum logic operation — just 0.000015%, or one error...

Dead Sea Scroll 4Q7, a fragment of the Book of Genesis found in Wadi Qumran Cave 4. Image credit: KetefHinnomFan.

Some Dead Sea Scrolls Older than Previously Thought, AI Analysis Suggests

Jun 4, 2025 by News Staff

Determining by means of paleography the chronology of ancient handwritten manuscripts such as the Dead Sea Scrolls is essential for reconstructing the...

Word rain of environment- and health-related GPT-fabricated, questionable full-text papers. Image credit: Haider et al., doi: 10.37016/mr-2020-156.

GPT-Fabricated Scientific Papers Flood Google Scholar, Scientists Say

Jan 14, 2025 by News Staff

In a new study published in the Harvard Kennedy School Misinformation Review, researchers from the University of Borås, Lund University and the Swedish...

A film a few atoms thick of non-crystalline niobium phosphide conducts better through the surface to make the material, as a whole, a better conductor. Image credit: Il-Kwon Oh / Asir Khan.

Researchers Create Ultrathin Niobium-Phosphide Conductor for Nanoelectronics

Jan 9, 2025 by News Staff

Niobium phosphide can conduct electricity better than copper in films that are only a few atoms thick; moreover, these films can be created and deposited...

Dayan et al. found that although large language models show remarkable proficiency in several cognitive domains, they show notable deficits in visuospatial and executive functions, akin to mild cognitive impairment in humans.

Almost All Leading Large Language Models Show Signs of Mild Cognitive Impairment: Study

Dec 20, 2024 by News Staff

With the exception of ChatGPT 4o, almost all publicly available large language models subjected to a test called the Montreal Cognitive Assessment (MoCA)...

The frequency of emoji use varies across genders and relationship types. Image credit: Pete Linforth.

Study: People with Higher Emotional Intelligence Use Emojis More Frequently

Dec 5, 2024 by News Staff

Higher emotional intelligence is linked to more emoji use with friends, while avoidant attachment is associated with less emoji use with friends and dating...

Bottom-up nano- to micro-fabrication is crucial in modern electronics and optics. Conventional multi-scale array fabrication techniques, however, are facing challenges in reconciling the contradiction between the pursuit of better device performance and lowering the fabrication cost and/or energy consumption. Chang et al. introduce a facile mixed-metal array fabrication method based on guided self-assembly of polymerizing organometallic adducts derived from the passivating oxides of a ternary liquid metal to create mixed metal wires. Image credit: Julia Chang.

Researchers Demonstrate New Technique for Self-Assembling Electronics

Dec 5, 2024 by News Staff

The new proof-of-concept work, published in the journal Materials Horizons, paves the way for self-assembling more complex electronic devices without relying...

The 5D memory crystal. Image credit: University of Southampton.

Researchers Store Entire Human Genome on ‘Everlasting’ 5D Memory Crystal

Sep 20, 2024 by News Staff

Unlike other data storage formats that degrade over time, 5D memory crystals can store up to 360 terabytes of information without loss for billions of...

Large Language Models Pose No Existential Threat to Humanity, Scientists Say

Aug 13, 2024 by News Staff

ChatGPT and other large language models (LLMs), comprising billions of parameters and pre-trained on extensive web-scale corpora, have been claimed to...

In this image, the person on the left (Scarlett Johansson) is real, while the person on the right is AI-generated. Their eyeballs are depicted underneath their faces. The reflections in the eyeballs are consistent for the real person, but incorrect (from a physics point of view) for the fake person. Image credit: Adejumoke Owolabi / CC BY 4.0.

Scientists Use Astronomy Techniques to Detect Deepfakes

Jul 22, 2024 by News Staff

Want to spot a deepfake? Look for the stars in their eyes, says a team of astronomers at the University of Hull. They suggest that AI-generated fakes can...

Schneiders et al. suggest it takes more than a carefully designed robot to care for your cat, the environment in which they operate is also vital, as well as human interaction. Image credit: Schneiders et al., doi: 10.1145/3613904.3642115.

Would You Trust Robot to Look After Your Cat?

May 14, 2024 by News Staff

Scientists from the University of Nottingham and artists from Blast Theory have created Cat Royale, a multispecies world centered around a bespoke enclosure...

Large language models and other AI systems have already learned, from their training, the ability to deceive via techniques such as manipulation, sycophancy, and cheating the safety test.

Large Language Models and Other AI Systems are Already Capable of Deceiving Humans, Scientists Say

May 14, 2024 by News Staff

In a new review paper published in the journal Patterns, researchers argue that a range of current AI systems have learned how to deceive humans. They...

Soltoggio et al. say we are set to see the emergence of Collective AI, where numerous artificial intelligence units, each capable of continuously acquiring new knowledge and skills, form a network to share information with each other.

AI’s Future is Similar to that of Star Trek’s Borg, Scientists Say

Mar 24, 2024 by News Staff

In a new paper in the journal Nature Machine Intelligence, leading computer scientists from around the world review recent machine learning advances converging...

Vasas et al. show 3 male orange sulphurs (Colias eurytheme). These butterflies display strong angle-dependent UV iridescence on the dorsal side of their wings. The UV-iridescent portions appear more orange to the human observer than the otherwise yellow wings (see human-visible inset). The avian false color depiction (main image) is based on the coordinates of the receptor noise-limited (RNL) opponent space. Coordinates in this space represent differences in an animal’s photoreceptor responses, and the distances between the coordinates approximate the perceptual distances between the colors. The inset color key illustrates human-visible colors around the perimeter of a circle colored according to RNL coordinates, where UV colors start at the epicenter and mix with other colors. In this depiction, the UV-iridescent portions appear purple, because they most strongly reflect in the ultraviolet and the red part of the spectrum. The ventral part of the wings, visible when the animals are in their resting position, are depicted as gray-brown, just like the leaves, as they are close to the achromatic point. Image credit: Vasas et al., doi: 10.1371/journal.pbio.3002444.

Novel Camera System and Software Records Natural Animal-View Videos with Over 90% Accuracy

Jan 24, 2024 by News Staff

A new camera system and associated software package allows scientists to produce videos that accurately replicate the colors that different animals see...

The tongue color is a mirror that reflects the health status of a patient. It could be used in diagnosing some diseases without any physical contact with the patient, which facilitates the process of primary healthcare. Image credit: Gerd Altmann / Sci.News.

New Diagnostic System Based on Tongue Color Analysis Correctly Diagnoses Diseases in 94% of Cases

Oct 24, 2023 by News Staff

Tongue diagnostic systems are fast gaining traction due to an increase in remote health monitoring worldwide, and new research from the University of South...

Problematic online usage and Internet addiction lack a refined classification. Using 796 questionnaires, Stangl et al. are the first to reveal a continuum ranging from casual users (i.e. older users) to addicts (i.e. digital natives) acknowledging five different severity levels of problematic online usage to Internet addiction. Image credit: Gerd Altmann.

Scientists Introduce New Internet Addiction Spectrum

Oct 2, 2023 by News Staff

Researchers from the University of Surrey and elsewhere have found that young people (24 years and youger) spend an average of six hours a day online;...