Welcome to the Nexus of Ethics, Psychology, Morality, Philosophy and Health Care

Welcome to the nexus of ethics, psychology, morality, technology, health care, and philosophy
Showing posts with label Evolution. Show all posts
Showing posts with label Evolution. Show all posts

Friday, March 1, 2024

AI needs the constraints of the human brain

Danyal Akarca
Originally posted 30 Jan 24

Here is an excerpt:

So, evolution shapes systems that are capable of solving competing problems that are both internal (e.g., how to expend energy) and external (e.g., how to act to survive), but in a way that can be highly efficient, in many cases elegant, and often surprising. But how does this evolutionary story of biological intelligence contrast with the current paradigm of AI?

In some ways, quite directly. Since the 50s, neural networks were developed as models that were inspired directly from neurons in the brain and the strength of their connections, in addition to many successful architectures of the past being directly motivated by neuroscience experimentation and theory. Yet, AI research in the modern era has occurred with a significant absence of thought of intelligent systems in nature and their guiding principles. Why is this? There are many reasons. But one is that the exponential growth of computing capabilities, enabled by increases of transistors on integrated circuits (observed since the 1950s, known as Moore’s Law), has permitted AI researchers to leverage significant improvements in performance without necessarily requiring extraordinarily elegant solutions. This is not to say that modern AI algorithms are not widely impressive – they are. It is just that the majority of the heavy lifting has come from advances in computing power rather than their engineered design. Consequently, there has been relatively little recent need or interest from AI experts to look to the brain for inspiration.

But the tide is turning. From a hardware perspective, Moore’s law will not continue ad infinitum (at 7 nanometers, transistor channel lengths are now nearing fundamental limits of atomic spacing). We will therefore not be able to leverage ever improving performance delivered by increasingly compact microprocessors. It is likely therefore that we will require entirely new computing paradigms, some of which may be inspired by the types of computations we observe in the brain (the most notable being neuromorphic computing). From a software and AI perspective, it is becoming increasingly clear that – in part due to the reliance on increases to computational power – the AI research field will need to refresh its conceptions as to what makes systems intelligent at all. For example, this will require much more sophisticated benchmarks of what it means to perform at human or super-human performance. In sum, the field will need to form a much richer view of the possible space of intelligent systems, and how artificial models can occupy different places in that space.

Key Points:
  • Evolutionary pressures: Efficient, resource-saving brains are advantageous for survival, leading to optimized solutions for learning, memory, and decision-making.
  • AI's reliance on brute force: Modern AI often achieves performance through raw computing power, neglecting principles like energy efficiency.
  • Shifting AI paradigm: Moore's Law's end and limitations in conventional AI call for exploration of new paradigms, potentially inspired by the brain.
  • Neurobiology's potential: Brain principles like network structure, local learning, and energy trade-offs can inform AI design for efficiency and novel functionality.
  • Embodied AI with constraints: Recent research incorporates space and communication limitations into AI models, leading to features resembling real brains and potentially more efficient information processing.

Saturday, January 13, 2024

Consciousness does not require a self

James Coook
Originally published 14 DEC 23

Here is an excerpt:

Beyond the neuroscientific study of consciousness, phenomenological analysis also reveals the self to not be the possessor of experience. In mystical experiences induced by meditation or psychedelics, individuals typically enter a mode of experience in which the psychological self is absent, yet consciousness remains. While this is not the default state of the mind, the presence of consciousness in the absence of a self shows that consciousness is not dependent on an experiencing subject. What is consciousness if not a capacity of an experiencing subject? Such an experience reveals consciousness to consist of a formless awareness at its core, an empty space in which experience arises, including the experience of being a self. The self does not possess consciousness, consciousness is the experiential space in which the image of a psychological self can appear. This mode of experience can be challenging to conceptualise but is very simple when experienced – it is a state of simple appearances arising without the extra add-on of a psychological self inspecting them.

We can think of a conscious system as a system that is capable of holding beliefs about the qualitative character of the world. We should not think of belief here as referring to complex conceptual beliefs, such as believing that Paris is the capital of France, but as the simple ability to hold that the world is a certain way. You do this when you visually perceive a red apple in front of you, the experience is one of believing the apple to exist with all of its qualities such as roundness and redness. This way of thinking is in line with the work of Immanuel Kant, who argued that we never come to know reality as it is but instead only experience phenomenal representations of reality [9]. We are not conscious of the world as it is, but as we believe it to be.

Here is my take:

For centuries, we've assumed consciousness and the sense of self are one and the same. This article throws a wrench in that assumption, proposing that consciousness can exist without a self. Imagine experiencing sights, sounds, and sensations without the constant "me" narrating it all. That's what "selfless consciousness" means – raw awareness untouched by self-reflection.

The article then posits that our familiar sense of self, complete with its stories and memories, isn't some fundamental truth but rather a clever prediction concocted by our brains. This "predicted self" helps us navigate the world and interact with others, but it's not necessarily who we truly are.

Decoupling consciousness from the self opens a Pandora's box of possibilities. We might find consciousness in unexpected places, like animals or even artificial intelligence. Understanding brain function could shift dramatically, and our very notions of identity, free will, and reality might need a serious rethink. This is a bold new perspective on what it means to be conscious, and its implications are quite dramatic.

Saturday, June 24, 2023

The Darwinian Argument for Worrying About AI

Dan Hendrycks
Originally posted 31 May 23

Here is an excerpt:

In the biological realm, evolution is a slow process. For humans, it takes nine months to create the next generation and around 20 years of schooling and parenting to produce fully functional adults. But scientists have observed meaningful evolutionary changes in species with rapid reproduction rates, like fruit flies, in fewer than 10 generations. Unconstrained by biology, AIs could adapt—and therefore evolve—even faster than fruit flies do.

There are three reasons this should worry us. The first is that selection effects make AIs difficult to control. Whereas AI researchers once spoke of “designing” AIs, they now speak of “steering” them. And even our ability to steer is slipping out of our grasp as we let AIs teach themselves and increasingly act in ways that even their creators do not fully understand. In advanced artificial neural networks, we understand the inputs that go into the system, but the output emerges from a “black box” with a decision-making process largely indecipherable to humans.

Second, evolution tends to produce selfish behavior. Amoral competition among AIs may select for undesirable traits. AIs that successfully gain influence and provide economic value will predominate, replacing AIs that act in a more narrow and constrained manner, even if this comes at the cost of lowering guardrails and safety measures. As an example, most businesses follow laws, but in situations where stealing trade secrets or deceiving regulators is highly lucrative and difficult to detect, a business that engages in such selfish behavior will most likely outperform its more principled competitors.

Selfishness doesn’t require malice or even sentience. When an AI automates a task and leaves a human jobless, this is selfish behavior without any intent. If competitive pressures continue to drive AI development, we shouldn’t be surprised if they act selfishly too.

The third reason is that evolutionary pressure will likely ingrain AIs with behaviors that promote self-preservation. Skeptics of AI risks often ask, “Couldn’t we just turn the AI off?” There are a variety of practical challenges here. The AI could be under the control of a different nation or a bad actor. Or AIs could be integrated into vital infrastructure, like power grids or the internet. When embedded into these critical systems, the cost of disabling them may prove too high for us to accept since we would become dependent on them. AIs could become embedded in our world in ways that we can’t easily reverse. But natural selection poses a more fundamental barrier: we will select against AIs that are easy to turn off, and we will come to depend on AIs that we are less likely to turn off.

These strong economic and strategic pressures to adopt the systems that are most effective mean that humans are incentivized to cede more and more power to AI systems that cannot be reliably controlled, putting us on a pathway toward being supplanted as the earth’s dominant species. There are no easy, surefire solutions to our predicament.

Monday, June 19, 2023

On the origin of laws by natural selection

DeScioli, P.
Evolution and Human Behavior
Volume 44, Issue 3, May 2023, Pages 195-209


Humans are lawmakers like we are toolmakers. Why do humans make so many laws? Here we examine the structure of laws to look for clues about how humans use them in evolutionary competition. We will see that laws are messages with a distinct combination of ideas. Laws are similar to threats but critical differences show that they have a different function. Instead, the structure of laws matches moral rules, revealing that laws derive from moral judgment. Moral judgment evolved as a strategy for choosing sides in conflicts by impartial rules of action—rather than by hierarchy or faction. For this purpose, humans can create endless laws to govern nearly any action. However, as prolific lawmakers, humans produce a confusion of contradictory laws, giving rise to a perpetual battle to control the laws. To illustrate, we visit some of the major conflicts over laws of violence, property, sex, faction, and power.


Moral rules are not for cooperation

We have briefly summarized the  major divisions and operations of moral judgment. Why then did humans evolve such elaborate powers of the mind devoted to moral rules? What is all this rule making for?

One common opinion is that moral rules are for cooperation. That is, we make and enforce a moral code in order to cooperate more effectively with other people. Indeed, traditional  theories beginning with Darwin assume that morality is  the  same  as cooperation. These theories  successfully explain many forms of cooperation, such as why humans and other  animals  care  for  offspring,  trade  favors,  respect  property, communicate  honestly,  and  work  together  in  groups.  For  instance, theories of reciprocity explain why humans keep records of other people’s deeds in the form of reputation, why we seek partners who are nice, kind, and generous, why we praise these virtues, and why we aspire to attain them.

However, if we look closely, these theories explain cooperation, not moral  judgment.  Cooperation pertains  to our decisions  to  benefit  or harm someone, whereas moral judgment pertains to  our judgments of someone’s action  as right or  wrong. The difference  is crucial because these  mental  faculties  operate  independently  and  they  evolved  separately. For  instance,  people can  use moral judgment  to cooperate but also to cheat, such as a thief who hides the theft because they judge it to be  wrong, or a corrupt leader who invents a  moral rule  that forbids criticism of the leader. Likewise, people use moral judgment to benefit others  but  also  to  harm  them, such  as falsely  accusing an enemy of murder to imprison them. 

Regarding  their  evolutionary  history, moral  judgment is  a  recent adaptation while cooperation is ancient and widespread, some forms as old  as  the origins  of  life and  multicellular  organisms.  Recalling our previous examples, social animals like gorillas, baboons, lions, and hyenas cooperate in numerous ways. They care for offspring, share food, respect property, work together in teams, form reputations,  and judge others’ characters as nice or nasty. But these species do not communicate rules of action, nor do they learn, invent, and debate the rules. Like language, moral judgment  most likely evolved  recently in the  human lineage, long after complex forms of cooperation. 

From the Conclusion

Having anchored ourselves to concrete laws, we next asked, What are laws for? This is the central question for  any mental power because it persists only  by aiding an animal in evolutionary competition.  In this search,  we  should  not  be  deterred  by  the  magnificent creativity  and variety of laws. Some people suppose that natural selection could impart no more than  a  few fixed laws in  the  human mind, but there  are  no grounds for this supposition. Natural selection designed all life on Earth and its creativity exceeds our own. The mental adaptations of animals outperform our best computer programs on routine tasks such as loco-motion and vision. Why suppose that human laws must be far simpler than, for instance, the flight controllers in the brain of a hummingbird? And there are obvious counterexamples. Language is a complex  adaptation but this does not mean that humans speak just a few sentences. Tool use comes from mental adaptations including an intuitive theory of physics, and again these abilities do not limit but enable the enormous variety of tools.

Sunday, May 14, 2023

Consciousness begins with feeling, not thinking

A. Damasio & H. Dimasio
Originally posted 20 APR 23

Please pause for a moment and notice what you are feeling now. Perhaps you notice a growing snarl of hunger in your stomach or a hum of stress in your chest. Perhaps you have a feeling of ease and expansiveness, or the tingling anticipation of a pleasure soon to come. Or perhaps you simply have a sense that you exist. Hunger and thirst, pain, pleasure and distress, along with the unadorned but relentless feelings of existence, are all examples of ‘homeostatic feelings’. Homeostatic feelings are, we argue here, the source of consciousness.

In effect, feelings are the mental translation of processes occurring in your body as it strives to balance its many systems, achieve homeostasis, and keep you alive. In a conventional sense feelings are part of the mind and yet they offer something extra to the mental processes. Feelings carry spontaneously conscious knowledge concerning the current state of the organism as a result of which you can act to save your life, such as when you respond to pain or thirst appropriately. The continued presence of feelings provides a continued perspective over the ongoing body processes; the presence of feelings lets the mind experience the life process along with other contents present in your mind, namely, the relentless perceptions that collect knowledge about the world along with reasonings, calculations, moral judgments, and the translation of all these contents in language form. By providing the mind with a ‘felt point of view’, feelings generate an ‘experiencer’, usually known as a self. The great mystery of consciousness in fact is the mystery behind the biological construction of this experiencer-self.

In sum, we propose that consciousness is the result of the continued presence of homeostatic feelings. We continuously experience feelings of one kind or another, and feelings naturally tell each of us, automatically, not only that we exist but that we exist in a physical body, vulnerable to discomfort yet open to countless pleasures as well. Feelings such as pain or pleasure provide you with consciousness, directly; they provide transparent knowledge about you. They tell you, in no uncertain terms, that you exist and where you exist, and point to what you need to do to continue existing – for example, treating pain or taking advantage of the well-being that came your way. Feelings illuminate all the other contents of mind with the light of consciousness, both the plain events and the sublime ideas. Thanks to feelings, consciousness fuses the body and mind processes and gives our selves a home inside that partnership.

That consciousness should come ‘down’ to feelings may surprise those who have been led to associate consciousness with the lofty top of the physiological heap. Feelings have been considered inferior to reason for so long that the idea that they are not only the noble beginning of sentient life but an important governor of life’s proceedings may be difficult to accept. Still, feelings and the consciousness they beget are largely about the simple but essential beginnings of sentient life, a life that is not merely lived but knows that it is being lived.

Tuesday, March 14, 2023

What Happens When AI Has Read Everything?

Ross Anderson
The Atlantic
Originally posted 18 JAN 23

Here is an excerpt:

Ten trillion words is enough to encompass all of humanity’s digitized books, all of our digitized scientific papers, and much of the blogosphere. That’s not to say that GPT-4 will have read all of that material, only that doing so is well within its technical reach. You could imagine its AI successors absorbing our entire deep-time textual record across their first few months, and then topping up with a two-hour reading vacation each January, during which they could mainline every book and scientific paper published the previous year.

Just because AIs will soon be able to read all of our books doesn’t mean they can catch up on all of the text we produce. The internet’s storage capacity is of an entirely different order, and it’s a much more democratic cultural-preservation technology than book publishing. Every year, billions of people write sentences that are stockpiled in its databases, many owned by social-media platforms.

Random text scraped from the internet generally doesn’t make for good training data, with Wikipedia articles being a notable exception. But perhaps future algorithms will allow AIs to wring sense from our aggregated tweets, Instagram captions, and Facebook statuses. Even so, these low-quality sources won’t be inexhaustible. According to Villalobos, within a few decades, speed-reading AIs will be powerful enough to ingest hundreds of trillions of words—including all those that human beings have so far stuffed into the web.

And the conclusion:

If, however, our data-gorging AIs do someday surpass human cognition, we will have to console ourselves with the fact that they are made in our image. AIs are not aliens. They are not the exotic other. They are of us, and they are from here. They have gazed upon the Earth’s landscapes. They have seen the sun setting on its oceans billions of times. They know our oldest stories. They use our names for the stars. Among the first words they learn are flow, mother, fire, and ash.

Sunday, November 6, 2022

‘Breakthrough’ finding shows how modern humans grow more brain cells than Neanderthals

Rodrigo Pérez Ortega
Originally posted 8 SEP 22

We humans are proud of our big brains, which are responsible for our ability to plan ahead, communicate, and create. Inside our skulls, we pack, on average, 86 billion neurons—up to three times more than those of our primate cousins. For years, researchers have tried to figure out how we manage to develop so many brain cells. Now, they’ve come a step closer: A new study shows a single amino acid change in a metabolic gene helps our brains develop more neurons than other mammals—and more than our extinct cousins, the Neanderthals.

The finding “is really a breakthrough,” says Brigitte Malgrange, a developmental neurobiologist at the University of Liège who was not involved in the study. “A single amino acid change is really, really important and gives rise to incredible consequences regarding the brain.”

What makes our brain human has been the interest of neurobiologist Wieland Huttner at the Max Planck Institute of Molecular Cell Biology and Genetics for years. In 2016, his team found that a mutation in the ARHGAP11B gene, found in humans, Neanderthals, and Denisovans but not other primates, caused more production of cells that develop into neurons. Although our brains are roughly the same size as those of Neanderthals, our brain shapes differ and we created complex technologies they never developed. So, Huttner and his team set out to find genetic differences between Neanderthals and modern humans, especially in cells that give rise to neurons of the neocortex. This region behind the forehead is the largest and most recently evolved part of our brain, where major cognitive processes happen.

The team focused on TKTL1, a gene that in modern humans has a single amino acid change—from lysine to arginine—from the version in Neanderthals and other mammals. By analyzing previously published data, researchers found that TKTL1 was mainly expressed in progenitor cells called basal radial glia, which give rise to most of the cortical neurons during development.

Thursday, September 22, 2022

Freezing revisited: coordinated autonomic and central optimization of threat coping

Roelofs, K., Dayan, P. 
Nat Rev Neurosci 23, 568–580 (2022).


Animals have sophisticated mechanisms for coping with danger. Freezing is a unique state that, upon threat detection, allows evidence to be gathered, response possibilities to be previsioned and preparations to be made for worst-case fight or flight. We propose that — rather than reflecting a passive fear state — the particular somatic and cognitive characteristics of freezing help to conceal overt responses, while optimizing sensory processing and action preparation. Critical for these functions are the neurotransmitters noradrenaline and acetylcholine, which modulate neural information processing and also control the sympathetic and parasympathetic branches of the autonomic nervous system. However, the interactions between autonomic systems and the brain during freezing, and the way in which they jointly coordinate responses, remain incompletely explored. We review the joint actions of these systems and offer a novel computational framework to describe their temporally harmonized integration. This reconceptualization of freezing has implications for its role in decision-making under threat and for psychopathology.

Conclusions and future directions

Considering the post encounter threat state from neural, psychological and computational perspectives has shown how the most obvious external characteristic of this state — a particular form of active freezing arising from co-activation of the normally opposed sympathetic and parasympathetic branches of the ANS — could have various advantages from the viewpoints of both information processing and fast Pavlovian or instrumental action. Descending control of this state is quite well understood, and the potential benefits of expending effort on enhancing unbiased, bottom-up, sensory processing and engaging in planning are easy to observe. However, the roles of ascending neuromodulators in engaging these forms of appropriate information processing are less clear.  Certainly, various of the modes of action of ACh and NA in the CNS are in a position to achieve some of this; but much remains to be discovered by precisely recording and manipulating the candidate circuits within the timeframes of the detection, evaluation and action stages.

One important source of ideas is evolutionary theory. For instance, the polyvagal theory of the phylogeny of the ANS suggests that it progressed in three stages. The first, associated with an unmyelinated vagus nerve, allowed metabolic activity to be depressed in response to threat and also controlled aspects of digestion. The second stage was associated with the sympathetic nervous system, which organized energized behaviour for fight or flight. The third stage was associated with a myelinated vagus nerve and allowed for more flexible and sophisticated responding. It has been suggested that the last stage is particularly involved in the evolution of somatic regulation in a social context; but the evolutionary layering of the competition and cooperation between the inhibitory and activating aspects of the different branches of the ANS is notable. It would be interesting to understand the parallel evolution of cholinergic and noradrenergic neuromodulation in the CNS. 

Note: We are primates subject to the principles of biology and evolution.

Sunday, August 21, 2022

Medial and orbital frontal cortex in decision-making and flexible behavior

Klein-Flügge, M. C., Bongioanni, A., & 
Rushworth, M. F. (2022).


The medial frontal cortex and adjacent orbitofrontal cortex have been the focus of investigations of decision-making, behavioral flexibility, and social behavior. We review studies conducted in humans, macaques, and rodents and argue that several regions with different functional roles can be identified in the dorsal anterior cingulate cortex, perigenual anterior cingulate cortex, anterior medial frontal cortex, ventromedial prefrontal cortex, and medial and lateral parts of the orbitofrontal cortex. There is increasing evidence that the manner in which these areas represent the value of the environment and specific choices is different from subcortical brain regions and more complex than previously thought. Although activity in some regions reflects distributions of reward and opportunities across the environment, in other cases, activity reflects the structural relationships between features of the environment that animals can use to infer what decision to take even if they have not encountered identical opportunities in the past.


Neural systems that represent the value of the environment exist in many vertebrates. An extended subcortical circuit spanning the striatum, midbrain, and brainstem nuclei of mammals corresponds to these ancient systems. In addition, however, mammals possess several frontal cortical regions concerned with guidance of decision-making and adaptive, flexible behavior. Although these frontal systems interact extensively with these subcortical circuits, they make specific contributions to behavior and also influence behavior via other cortical routes. Some areas such as the ACC, which is present in a broad range of mammals, represent the distribution of opportunities in an environment over space and time, whereas other brain regions such as amFC and dmPFC have roles in representing structural associations and causal links between environmental features, including aspects of the social environment (Figure 8). Although the origins of these areas and their functions are traceable to rodents, they are especially prominent in primates. They make it possible not just to select choices on the basis of past experience of identical situations, but to make inferences to guide decisions in new scenarios.

Friday, June 3, 2022

Cooperation as a signal of time preferences

Lie-Panis, J., & André, J. (2021, June 23).


Many evolutionary models explain why we cooperate with non kin, but few explain why cooperative behavior and trust vary. Here, we introduce a model of cooperation as a signal of time preferences, which addresses this variability. At equilibrium in our model, (i) future-oriented individuals are more motivated to cooperate, (ii) future-oriented populations have access to a wider range of cooperative opportunities, and (iii) spontaneous and inconspicuous cooperation reveal stronger preference for the future, and therefore inspire more trust. Our theory sheds light on the variability of cooperative behavior and trust. Since affluence tends to align with time preferences, results (i) and (ii) explain why cooperation is often associated with affluence, in surveys and field studies. Time preferences also explain why we trust others based on proxies for impulsivity, and, following result (iii), why uncalculating, subtle and one-shot cooperators are deemed particularly trustworthy. Time preferences provide a powerful and parsimonious explanatory lens, through which we can better understand the variability of trust and cooperation.

From the Discussion Section

Trust depends on revealed time preferences

Result (iii) helps explain why we infer trustworthiness from traits which appear unrelated  to cooperation,  but  happen  to  predict  time  preferences.   We  trust known partners and strangers based on how impulsive we perceive them to be (Peetz & Kammrath, 2013; Righetti & Finkenauer, 2011); impulsivity being associated to both time preferences and cooperativeness in laboratory experiments (Aguilar-Pardo et al., 2013; Burks et al., 2009; Cohen et al., 2014; Martinsson et al., 2014; Myrseth et al., 2015; Restubog et al., 2010).  Other studies show we infer cooperative motivation from a wide variety of proxies for partner self-control, including indicators of their indulgence in harmless sensual pleasures (for a review see  Fitouchi et al.,  2021),  as well as proxies for environmental affluence (Moon et al., 2018; Williams et al., 2016).

Time preferences further offer a parsimonious explanation for why different forms of cooperation inspire more trust than others.  When probability of observation p or cost-benefit ratio r/c are small in our model, helpful behavior reveals large time horizon- and cooperators may be perceived as relatively genuine or disinterested.  We derive two different types of conclusion from this principle.  (Inconspicuous and/or spontaneous cooperation)

Friday, February 25, 2022

Public Deliberation about Gene Editing in the Wild

M. K. Gusmano, E. Kaebnick, et al. (2021).
Hastings Center Report
10.1002/hast.1318, 51, S2, (S34-S41).


Genetic editing technologies have long been used to modify domesticated nonhuman animals and plants. Recently, attention and funding have also been directed toward projects for modifying nonhuman organisms in the shared environment—that is, in the “wild.” Interest in gene editing nonhuman organisms for wild release is motivated by a variety of goals, and such releases hold the possibility of significant, potentially transformative benefit. The technologies also pose risks and are often surrounded by a high uncertainty. Given the stakes, scientists and advisory bodies have called for public engagement in the science, ethics, and governance of gene editing research in nonhuman organisms. Most calls for public engagement lack details about how to design a broad public deliberation, including questions about participation, how to structure the conversations, how to report on the content, and how to link the deliberations to policy. We summarize the key design elements that can improve broad public deliberations about gene editing in the wild.

Here is the gist of the paper:

We draw on interdisciplinary scholarship in bioethics, political science, and public administration to move forward on this knot of conceptual, normative, and practical problems. When is broad public deliberation about gene editing in the wild necessary? And when it is required, how should it be done? These questions lead to a suite of further questions about, for example, the rationale and goals of deliberation, the features of these technologies that make public deliberation appropriate or inappropriate, the criteria by which “stakeholders” and “relevant publics” for these uses might be identified, how different approaches to public deliberation map onto the challenges posed by the technologies, how the topic to be deliberated upon should be framed, and how the outcomes of public deliberation can be meaningfully connected to policy-making.

Tuesday, January 18, 2022

MIT Researchers Just Discovered an AI Mimicking the Brain on Its Own

Eric James Beyer
Interesting Engineering
Originally posted 18 DEC 21

Here is an excerpt:

In the wake of these successes, Martin began to wonder whether or not the same principle could be applied to higher-level cognitive functions like language processing. 

“I said, let’s just look at neural networks that are successful and see if they’re anything like the brain. My bet was that it would work, at least to some extent.”

To find out, Martin and colleagues compared data from 43 artificial neural network language models against fMRI and ECoG neural recordings taken while subjects listened to or read words as part of a text. The AI models the group surveyed covered all the major classes of available neural network approaches for language-based tasks. Some of them were more basic embedding models like GloVe, which clusters semantically similar words together in groups. Others, like the models known as GPT and BERT, were far more complex. These models are trained to predict the next word in a sequence or predict a missing word within a certain context, respectively. 

“The setup itself becomes quite simple,” Martin explains. “You just show the same stimuli to the models that you show to the subjects [...]. At the end of the day, you’re left with two matrices, and you test if those matrices are similar.”

And the results? 

“I think there are three-and-a-half major findings here,” Schrimpf says with a laugh. “I say ‘and a half’ because the last one we still don’t fully understand.”

Machine learning that mirrors the brain

The finding that sticks out to Martin most immediately is that some of the models predict neural data extremely well. In other words, regardless of how good a model was at performing a task, some of them appear to resemble the brain’s cognitive mechanics for language processing. Intriguingly, the team at MIT identified the GPT model variants as the most brain-like out of the group they looked at.

Thursday, December 9, 2021

'Moral molecules’ – a new theory of what goodness is made of

Oliver Scott Curry and others
Originally posted 1 NOV 21

Here are two excerpts:

Research is converging on the idea that morality is a collection of rules for promoting cooperation – rules that help us work together, get along, keep the peace and promote the common good. The basic idea is that humans are social animals who have lived together in groups for millions of years. During this time, we have been surrounded by opportunities for cooperation – for mutually beneficial social interaction – and we have evolved and invented a range of ways of unlocking these benefits. These cooperative strategies come in different shapes and sizes: instincts, intuitions, inventions, institutions. Together, they motivate our cooperative behaviour and provide the criteria by which we evaluate the behaviour of others. And it is these cooperative strategies that philosophers and others have called ‘morality’.

This theory of ‘morality as cooperation’ relies on the mathematical analysis of cooperation provided by game theory – the branch of maths that is used to describe situations in which the outcome of one’s decisions depends on the decisions made by others. Game theory distinguishes between competitive ‘zero-sum’ interactions or ‘games’, where one player’s gain is another’s loss, and cooperative ‘nonzero-sum’ games, win-win situations in which both players benefit. What’s more, game theory tells us that there is not just one type of nonzero-sum game; there are many, with many different cooperative strategies for playing them. At least seven different types of cooperation have been identified so far, and each one explains a different type of morality.


Hence, seven types of cooperation explain seven types of morality: love, loyalty, reciprocity, heroism, deference, fairness and property rights. And so, according to this theory, it is morally good to: 1) love your family; 2) be loyal to your group; 3) return favours; 4) be heroic; 5) defer to superiors; 6) be fair; and 7) respect property. (And it is morally bad to: 1) neglect your family; 2) betray your group; 3) cheat; 4) be a coward; 5) disrespect authority; 6) be unfair; or 7) steal.) These morals are evolutionarily ancient, genetically distinct, psychologically discrete and cross-culturally universal.

The theory of ‘morality as cooperation’ explains, from first principles, many of the morals on those old lists. Some of the morals correspond to one of the basic types of cooperation (as in the case of courage), while others correspond to component parts of a basic type (as in the case of gratitude, which is a component of reciprocity).

Wednesday, December 8, 2021

Robot Evolution: Ethical Concerns

Eiban, A.E., Ellers, J, et al.
Front. Robot. AI, 03 November 2021


Rapid developments in evolutionary computation, robotics, 3D-printing, and material science are enabling advanced systems of robots that can autonomously reproduce and evolve. The emerging technology of robot evolution challenges existing AI ethics because the inherent adaptivity, stochasticity, and complexity of evolutionary systems severely weaken human control and induce new types of hazards. In this paper we address the question how robot evolution can be responsibly controlled to avoid safety risks. We discuss risks related to robot multiplication, maladaptation, and domination and suggest solutions for meaningful human control. Such concerns may seem far-fetched now, however, we posit that awareness must be created before the technology becomes mature.


Robot evolution is not science fiction anymore. The theory and the algorithms are available and robots are already evolving in computer simulations, safely limited to virtual worlds. In the meanwhile, the technology for real-world implementations is developing rapidly and the first (semi-) autonomously reproducing and evolving robots are likely to arrive within a decade (Hale et al., 2019; Buchanan et al., 2020). Current research in this area is typically curiosity-driven, but will increasingly become more application-oriented as evolving robot systems can be employed in hostile or inaccessible environments, like seafloors, rain-forests, ultra-deep mines or other planets, where they develop themselves “on the job” without the need for direct human oversight.

A key insight of this paper is that the practice of second order engineering, as induced by robot evolution, raises new issues outside the current discourse on AI and robot ethics. Our main message is that awareness must be created before the technology becomes mature and researchers and potential users should discuss how robot evolution can be responsibly controlled. Specifically, robot evolution needs careful ethical and methodological guidelines in order to minimize potential harms and maximize the benefits. Even though the evolutionary process is functionally autonomous without a “steering wheel” it still entails a necessity to assign responsibilities. This is crucial not only with respect to holding someone responsible if things go wrong, but also to make sure that people take responsibility for certain aspects of the process–without people taking responsibility, the process cannot be effectively controlled. Given the potential benefits and harms and the complicated control issues, there is an urgent need to follow up our ideas and further think about responsible robot evolution.

Tuesday, November 2, 2021

Our evolved intuitions about privacy aren’t made for this era

Joe Green & Azim Shariff
Originally published September 16, 2021

Here is an excerpt:

Our concern for privacy has its evolutionary roots in the need to maintain boundaries between the self and others, for safety and security. The motivation for personal space and territoriality is a common phenomenon within the animal kingdom. Among humans, this concern about regulating physical access is complemented by one about regulating informational access. The language abilities, complex social lives and long memories of human beings made protecting our social reputations almost as important as protecting our physical bodies. Norms about sexual privacy, for instance, are common across cultures and time periods. Establishing basic seclusion for secret trysts would have allowed for all the carnal benefits without the unwelcome reputational scrutiny.

Since protection and seclusion must be balanced with interaction, our privacy concern is tuned to flexibly respond to cues in our environment, helping to determine when and what and with whom we share our physical space and personal information. We reflexively lower our voices when strange or hostile interlopers come within earshot. We experience an uneasy creepiness when someone peers over our shoulder. We viscerally feel the presence of a crowd and the public scrutiny that comes with it.

However, just as the turtles’ light-orienting reflex was confounded by the glow of urban settlements, so too have our privacy reactions been confounded by technology. Cameras and microphones – with their superhuman sensory abilities – were challenging enough. But the migration of so much of our lives online is arguably the largest environmental shift in our species’ history with regard to privacy. And our evolved privacy psychology has not caught up. Consider how most people respond to the presence of others when they are in a crowd. Humans use a host of social cues to regulate how much distance they keep between themselves and others. These include facial expression, gaze, vocal quality, posture and hand gestures. In a crowd, such cues can produce an anxiety-inducing cacophony. Moreover, our hair-trigger reputation-management system – critical to keeping us in good moral standing within our group – can drive us into a delirium of self-consciousness.

However, there is some wisdom in this anxiety. Looking into the whites of another’s eyes anchors us within the social milieu, along with all of its attendant norms and expectations. As a result, we tread carefully. Our private thoughts generally remain just that – private, conveyed only to small, trusted groups or confined to our own minds. But as ‘social networks’ suddenly switched from being small, familiar, in-person groupings to online social media platforms connecting millions of users, things changed. Untethered from recognisable social cues such as crowding and proximity, thoughts better left for a select few found their way in front of a much wider array of people, many of whom do not have our best interests at heart. Online we can feel alone and untouchable when we are neither.

Consider, too, our intuitions about what belongs to whom. Ownership can be complicated from a legal perspective but, psychologically, it is readily inferred from an early age (as anyone with young children will have realised). This is achieved through a set of heuristics that provide an intuitive ‘folk psychology’ of ownership. First possession (who first possessed an object), labour investment (who made or modified an object), and object history (information about past transfer of ownership) are all cues that people reflexively use in attributing the ownership of physical things – and consequently, the right to open, inspect or enter them.

Tuesday, October 19, 2021

Why Empathy Is Not a Reliable Source of Information in Moral Decision Making

Decety, J. (2021).
Current Directions in Psychological Science. 


Although empathy drives prosocial behaviors, it is not always a reliable source of information in moral decision making. In this essay, I integrate evolutionary theory, behavioral economics, psychology, and social neuroscience to demonstrate why and how empathy is unconsciously and rapidly modulated by various social signals and situational factors. This theoretical framework explains why decision making that relies solely on empathy is not ideal and can, at times, erode ethical values. This perspective has social and societal implications and can be used to reduce cognitive biases and guide moral decisions.

From the Conclusion

Empathy can encourage overvaluing some people and ignoring others, and privileging one over many. Reasoning is therefore essential to filter and evaluate emotional responses that guide moral decisions. Understanding the ultimate causes and proximate mechanisms of empathy allows characterization of the kinds of signals that are prioritized and identification of situational factors that exacerbate empathic failure. Together, this knowledge is useful at a theoretical level, and additionally provides practical information about how to reframe situations to activate alternative evolved systems in ways that promote normative moral conduct compatible with current societal aspirations. This conceptual framework advances current understanding of the role of empathy in moral decision making and may inform efforts to correct personal biases. Becoming aware of one’s biases is not the most effective way to manage and mitigate them, but empathy is not something that can be ignored. It has an adaptive biological function, after all.

Sunday, October 17, 2021

The Cognitive Science of Technology

D. Stout
Trends in Cognitive Sciences
Available online 4 August 2021


Technology is central to human life but hard to define and study. This review synthesizes advances in fields from anthropology to evolutionary biology and neuroscience to propose an interdisciplinary cognitive science of technology. The foundation of this effort is an evolutionarily motivated definition of technology that highlights three key features: material production, social collaboration, and cultural reproduction. This broad scope respects the complexity of the subject but poses a challenge for theoretical unification. Addressing this challenge requires a comparative approach to reduce the diversity of real-world technological cognition to a smaller number of recurring processes and relationships. To this end, a synthetic perceptual-motor hypothesis (PMH) for the evolutionary–developmental–cultural construction of technological cognition is advanced as an initial target for investigation.

  • Evolutionary theory and paleoanthropological/archaeological evidence motivate a theoretical definition of technology as socially reproduced and elaborated behavior involving the manipulation and modification of objects to enact changes in the physical environment.
  • This definition helps to resolve or obviate ongoing controversies in the anthropological, neuroscientific, and psychological literature relevant to technology.
  • A review of evidence from across these disciplines reveals that real-world technologies are diverse in detail but unified by the underlying demands and dynamics of material production. This creates opportunities for meaningful synthesis using a comparative method.
  • A ‘perceptual‐motor hypothesis’ proposes that technological cognition is constructed on biocultural evolutionary and developmental time scales from ancient primate systems for sensorimotor prediction and control.

Tuesday, July 6, 2021

On the Origins of Diversity in Social Behavior

Young, L.J. & Zhang, Q.
Japanese Journal of Animal Psychology


Here we discuss the origins of diversity in social behavior by highlighting research using the socially monogamous prairie vole. Prairie voles display a rich social behavioral repertoire involving pair bonding and consoling behavior that are not observed in typical laboratory species. Oxytocin and vasopressin play critical roles in regulating pair bonding and consoling behavior. Oxytocin and vasopressin receptors show remarkable diversity in expression patterns both between and within species. Receptor expression patterns are associated with species differences in social behaviors. Variations in receptor genes have been linked to individual variation in expression patterns. We propose that "evolvability" in the oxytocin and vasopressin receptor genes allows for the repurposing of ancient maternal and territorial circuits to give rise to novel social behaviors such as pair bonding, consoling and selective aggression. We further propose that the evolvability of these receptor genes is due to their transcriptional sensitivity to genomic variation. This model provides a foundation for investigating the molecular mechanisms giving rise to the remarkable diversity in social behaviors found in vertebrates.

While this hypothesis remains to be tested, we believe this transcriptional flexibility is key to the origin of diversity in social behavior, and enables rapid social behavioral adaptation through natural selection, and
contributes to the remarkable diversity in social and reproductive behaviors in the animal kingdom.

Sunday, June 20, 2021

Artificial intelligence research may have hit a dead end

Thomas Nail
Originally published 30 April 21

Here is an excerpt:

If it's true that cognitive fluctuations are requisite for consciousness, it would also take time for stable frequencies to emerge and then synchronize with one another in resting states. And indeed, this is precisely what we see in children's brains when they develop higher and more nested neural frequencies over time.

Thus, a general AI would probably not be brilliant in the beginning. Intelligence evolved through the mobility of organisms trying to synchronize their fluctuations with the world. It takes time to move through the world and learn to sync up with it. As the science fiction author Ted Chiang writes, "experience is algorithmically incompressible." 

This is also why dreaming is so important. Experimental research confirms that dreams help consolidate memories and facilitate learning. Dreaming is also a state of exceptionally playful and freely associated cognitive fluctuations. If this is true, why should we expect human-level intelligence to emerge without dreams? This is why newborns dream twice as much as adults, if they dream during REM sleep. They have a lot to learn, as would androids.

In my view, there will be no progress toward human-level AI until researchers stop trying to design computational slaves for capitalism and start taking the genuine source of intelligence seriously: fluctuating electric sheep.

Monday, June 7, 2021

Science Skepticism Across 24 Countries

Rutjens, B. T., et al., (2021). 
Social Psychological and Personality Science. 


Efforts to understand and remedy the rejection of science are impeded by lack of insight into how it varies in degree and in kind around the world. The current work investigates science skepticism in 24 countries (N = 5,973). Results show that while some countries stand out as generally high or low in skepticism, predictors of science skepticism are relatively similar across countries. One notable effect was consistent across countries though stronger in Western, Educated, Industrialized, Rich, and Democratic (WEIRD) nations: General faith in science was predicted by spirituality, suggesting that it, more than religiosity, may be the ‘enemy’ of science acceptance. Climate change skepticism was mainly associated with political conservatism especially in North America. Other findings were observed across WEIRD and non-WEIRD nations: Vaccine skepticism was associated with spirituality and scientific literacy, genetic modification skepticism with scientific literacy, and evolution skepticism with religious orthodoxy. Levels of science skepticism are heterogeneous across countries, but predictors of science skepticism are heterogeneous across domains.

From the Discussion

Indeed, confirming previous results obtained in the Netherlands (Rutjens & van der Lee, 2020)—and providing strong support for Hypothesis 6—the current data speak to the crucial role of spirituality in fostering low faith in science, more generally, beyond its domain-specific effects on vaccine skepticism. This indicates that the negative impact of spirituality on faith in science represents a cross-national phenomenon that is more generalizable than might be expected based on the large variety (Muthukrishna et al., 2020) of countries included here. A possible explanation for the robustness of this effect may lie in the inherent irreconcilability of the intuitive epistemology of a spiritual belief system with science (Rutjens & van der Lee, 2020). (If so, then we might look at a potentially much larger problem that extends beyond spirituality and applies more generally to “post-truth” society, in which truth and perceptions of reality may be based on feelings rather than facts; Martel et al., 2020; Rutjens & Brandt, 2018.) However, these results do not mean that traditional religiosity as a predictor of science skepticism (McPhetres & Zuckermann, 2018; Rutjens, Heine, et al., 2018; Rutjens, Sutton, & van der Lee, 2018) has now become irrelevant: Not only did religious orthodoxy significantly contribute to low faith in science, it was also found to be a very consistent cross-national predictor of evolution skepticism (but not of other forms of science skepticism included in the study).