Lifeboat Foundation

Jailbroken large language models (LLMs) and generative AI chatbots — the kind any hacker can access on the open Web — are capable of providing in-depth, accurate instructions for carrying out large-scale acts of destruction, including bio-weapons attacks.

An alarming new study from RAND, the US nonprofit think tank, offers a canary in the coal mine for how bad actors might weaponize this technology in the (possibly near) future.

In an experiment, experts asked an uncensored LLM to plot out theoretical biological weapons attacks against large populations. The AI algorithm was detailed in its response and more than forthcoming in its advice on how to cause the most damage possible, and acquire relevant chemicals without raising suspicion.

“For AI to be motivated towards a goal, it must know what it wants.”

The possible board combinations in a game of Go are more than the number of atoms in the known universe, but it’s still a finite number. In the real world, there are infinite possibilities for what might happen next, and uncertainty is rampant. How realistic, then, is AGI?

A recent research paper published in Frontiers in Ecology and Evolution explores obstacles toward AGI. Biological systems with degrees of general intelligence — organisms ranging from the humble microbes to the humans reading this — are capable of improvising to meet their goals. What prevents AI from improvising?

Scanning tunnelling microscopy images of simple glycoconjugates and glycosaminoglycans and their corresponding structures.

Scanning tunnelling microscopy has enabled researchers to directly image important sugar molecules attached to lipids and proteins. The experiments provide a picture at the single-molecule level of the sequences and locations of glycans bound to important biomolecules, offering new insight into the role they play in biology.

WASHINGTON, Oct 16 (Reuters) — When British naturalist Charles Darwin sketched out his theory of evolution in the 1,859 book “On the Origin of Species” — proposing that biological species change over time through the acquisition of traits that favor survival and reproduction — it provoked a revolution in scientific thought.

Now 164 years later, nine scientists and philosophers on Monday proposed a new law of nature that includes the biological evolution described by Darwin as a vibrant example of a much broader phenomenon, one that appears at the level of atoms, minerals, planetary atmospheres, planets, stars and more.

It holds that complex natural systems evolve to states of greater patterning, diversity and complexity.

Are humans progressing morally as well as materially? What does it mean to be human in the cosmos? On a new episode of ID the Future, we bring you the second half of a stimulating conversation between Dr. David Berlinski and host Eric Metaxas on the subject of Berlinski’s book Human Nature.

In Human Nature, Berlinski argues that the utopian view that humans are progressing toward evolutionary and technological perfection is wishful thinking. Men are not about to become like gods. “I’m a strong believer in original sin,” quips Berlinski in his discussion with Metaxas. In other words, he believes not only that humans are fundamentally distinct from the rest of the biological world, but also that humans are prone to ignorance and depravity as well as wisdom and nobility. During this second half of their discussion, Berlinski and Metaxas compare and contrast the ideas of thinkers like psychologist Steven Pinker, author Christopher Hitchens, and physicist Steven Weinberg. The pair also spar gracefully over the implications of human uniqueness. Berlinski, though candid and self-critical, is unwilling to be pigeonholed. Metaxas, drawing his own conclusions about the role of mind in the universe, challenges Berlinski into moments of clarity with his usual charm. The result is an honest, probing, and wide-ranging conversation about the nature of science and the human condition. Download the podcast or listen to it here.

This is Part 2 of a two-part interview. If you missed it, listen to Part 1.

Recent developments in neuroscience and brain-inspired artificial intelligence have opened up new possibilities in understanding intelligence. Now, a research team led by Tianzi Jiang at the Institute of Automation of the Chinese Academy of Sciences has outlined the key components and properties of an innovative platform called the Digital Twin Brain, which could bridge the gap between biological and artificial intelligence and provide new insights into both. This research was published Sept. 22 in Intelligent Computing, a Science Partner Journal.

Network structure is something that biological and artificial intelligence have in common. Since the brain consists of biological networks, a digital model or “twin” of the brain built using artificial networks would allow researchers to feed knowledge about biological intelligence into the model. The ultimate goal is to “propel the development of artificial general intelligence and facilitate precision mental healthcare,” a feat calling for joint efforts from interdisciplinary scientists worldwide.

Using the Digital Twin Brain, researchers could explore the working mechanisms of the human brain by simulating and modulating the brain in different states for various cognitive tasks. For example, they could simulate how the brain functions properly in a resting state and how it malfunctions in disorders, or develop methods to shift it away from an undesirable state by modulating its activity.

‘Could the theory be wrong? Possibly. That is the point and the case for all theories,’ says Cronin. ‘But perhaps it is less wrong than our current understanding and it will help us understand the link between physics and biology through chemistry. We have to try and we think we are onto something.’

A Sharma et al, Nature., 2023, DOI: 10.1038/s41586-023–06600-9.

Life on Earth began from a single-celled microbe, while the rise to the multicellular world in which we live arose due a vital chemical process known as biomineralization, during which living organisms produce hardened mineralized tissue, such as skeletons. Not only did this phenomenon give rise to the plethora of body plans we see today, but it also had a major impact on the planet’s carbon cycle.

Fossil skeletons of cloudinids (Cloudina), tubular structures comprised of carbonate cones up to ~1.5cm in length, have been found in Tsau Khaeb National Park, Namibia, dating back to 551–550 million years ago in the Ediacaran (~635–538 million years ago). Dr. Fred Bowyer, from the University of Edinburgh, and colleagues aimed to use these fossils to define the location, timing and reason for why biomineralization initiated on Earth and the magnitude of its impact.

New research published in Earth and Planetary Science Letters combines sediment analysis with geochemical data in the form of carbon and oxygen isotopes (the same element with different atomic masses) from limestones in the Kliphoek Member, Nama Group. The research team suggest this rock was once deposited in a shallow sea during a lowstand before a period of transition to open marine conditions.

The anterior cingulate cortex (ACC) is believed to be involved in many cognitive processes, including linking goals to actions and tracking decision-relevant contextual information. ACC neurons robustly encode expected outcomes, but how this relates to putative functions of ACC remains unknown. Here, we approach this question from the perspective of population codes by analyzing neural spiking data in the ventral and dorsal banks of the ACC in two male monkeys trained to perform a stimulus-motor mapping task to earn rewards or avoid losses. We found that neural populations favor a low dimensional representational geometry that emphasizes the valence of potential outcomes while also facilitating the independent, abstract representation of multiple task-relevant variables. Valence encoding persisted throughout the trial, and realized outcomes were primarily encoded in a relative sense, such that cue valence acted as a context for outcome encoding. This suggests that the population coding we observe could be a mechanism that allows feedback to be interpreted in a context-dependent manner. Together, our results point to a prominent role for ACC in context setting and relative interpretation of outcomes, facilitated by abstract, or untangled, representations of task variables.

SIGNIFICANCE STATEMENT The ability to interpret events in light of the current context is a critical facet of higher-order cognition. The ACC is suggested to be important for tracking contextual information, whereas alternate views hold that its function is more related to the motor system and linking goals to appropriate actions. We evaluated these possibilities by analyzing geometric properties of neural population activity in monkey ACC when contexts were determined by the valence of potential outcomes and found that this information was represented as a dominant, abstract concept. Ensuing outcomes were then coded relative to these contexts, suggesting an important role for these representations in context-dependent evaluation. Such mechanisms may be critical for the abstract reasoning and generalization characteristic of biological intelligence.