The true self and decision-making capacity.

Toomey, J., Lewis, J., Hannikainen, I., & Earp, B. D.

(2024). The American Journal of Bioethics, in press.

Jennifer Hawkins (2024) offers two cases that challenge traditional accounts of decision-making capacity, according to which respect for a medical decision turns on an individual’s cognitive capacities at the time the decision is made (Hawkins 2024; Appelbaum and Grisso 1988). In each of her described cases (involving anorexia nervosa and grief, respectively), a patient makes a decision that—although instrumentally rational at the time—does not reflect the patient’s longer-term values due to being in a particular psychological state. Importantly, this state does not impair the patient’s cognition, but rather predisposes them to make a decision that conflicts with their own broader values, beliefs, or desires.

Under traditional understandings of decision-making capacity, the patient’s decision in either case must be followed by healthcare providers, insofar as it was made while in possession of the requisite cognitive abilities. But this, Hawkins suggests, is the wrong outcome. Although core cognitive capacities are necessary for decision-making capacity, they are not on her view sufficient. From her perspective, patients who clear the threshold of cognitive capacity are not entitled to have their decisions followed when there is good evidence they are making a serious prudential mistake while known to have a condition that makes people more likely than typical to make such mistakes.

A pdf is here.

Here are a few thoughts:

This article highlights a crucial topic regarding decision-making.

Traditionally, a patient's cognitive abilities have been the main focus for evaluating their decision-making capacity. This article challenges that notion by introducing the concept of a patient's "true self". It argues that even if a patient has the cognitive ability to make a decision, it shouldn't be automatically respected if it doesn't reflect their long-term values or identity. This emphasizes the importance of understanding a patient's underlying values and goals when making decisions about their care. The article aligns with the idea that people often judge medical decisions based on a patient's "true self," highlighting the complexity of determining a patient's true wishes.

Finally, the article confronts us with ethical dilemmas where respecting a patient's cognitive decision might conflict with their well-being. This underlines the need for a nuanced approach that considers both cognitive capacity and a patient's identity. In essence, this article encourages mental health professionals to move beyond following just cognitive protocols and strive for a more comprehensive understanding of their patients.

What makes full artificial agents morally different

Firt, E.

AI & Soc (2024).

https://doi.org/10.1007/s00146-024-01867-6

Abstract

In the research field of machine ethics, we commonly categorize artificial moral agents into four types, with the most advanced referred to as a full ethical agent, or sometimes a full-blown Artificial Moral Agent (AMA). This type has three main characteristics: autonomy, moral understanding and a certain level of consciousness, including intentional mental states, moral emotions such as compassion, the ability to praise and condemn, and a conscience. This paper aims to discuss various aspects of full-blown AMAs and presents the following argument: the creation of full-blown artificial moral agents, endowed with intentional mental states and moral emotions, and trained to align with human values, does not, by itself, guarantee that these systems will have human morality. Therefore, it is questionable whether they will be inclined to honor and follow what they perceive as incorrect moral values. we do not intend to claim that there is such a thing as a universally shared human morality, only that as there are different human communities holding different sets of moral values, the moral systems or values of the discussed artificial agents would be different from those held by human communities, for reasons we discuss in the paper.

The article is here.

Here are some thoughts:

This article discusses the complex implications of creating advanced artificial moral agents (AMAs) and the potential future coexistence of these agents with hybrid humans. The authors present a compelling argument that full-blown AMAs, despite being designed to align with human values, may not necessarily share or adhere to human morality.

The key insight here is that even if we create AMAs with autonomy, moral understanding, and consciousness, their moral systems may fundamentally differ from those of human communities. This challenges the assumption that equipping AI systems with human-like emotions and empathy will ensure they behave in ways we consider proper or controllable.

The authors raise an important question about the feasibility of aligning the values of sophisticated AMAs with human values, especially if these agents develop moral systems that are incomprehensible or foreign to us. This presents a significant challenge in terms of trust and control, as we cannot assume these agents will follow moral principles that contradict their own inner convictions.

The discussion on hybrid humans adds another layer of complexity to this ethical landscape. As humans integrate more with technology, their moral perspectives may shift, potentially becoming more sympathetic to the different moral values of full ethical agents. This evolution could lead to a scenario where the distinction between hybrid humans and AMAs becomes increasingly blurred.

From a clinical psychology perspective, this raises fascinating questions about the nature of morality, consciousness, and identity. How might the integration of artificial components into human beings affect their moral decision-making processes and emotional responses? As psychologists, we must consider the potential psychological impacts of such radical changes on individuals and society as a whole.

Furthermore, the ethical dilemmas presented here highlight the need for interdisciplinary collaboration in addressing these future challenges. Psychologists, ethicists, AI researchers, and policymakers must work together to navigate the complex terrain of artificial morality and its implications for human society.

In conclusion, this article underscores the importance of carefully considering the long-term consequences of creating advanced artificial moral agents. It challenges us to think deeply about the nature of morality, the potential divergence between human and artificial ethical systems, and how we might prepare for a future where the lines between human and artificial intelligence become increasingly blurred.

Moral agents as relational systems: The Contract-Based Model of Moral Cognition for AI

Vidal, L. M., Marchesi, S., Wykowska, A., & Pretus, C.

(2024, July 3)

 https://doi.org/10.31234/osf.io/52x74

Abstract

As artificial systems are becoming more prevalent in our daily lives, we should ensure that they make decisions that are aligned with human values. Utilitarian algorithms, which aim to maximize benefits and minimize harm fall short when it comes to human autonomy and fairness since it is insensitive to other-centered human preferences or how the burdens and benefits are distributed, as long as the majority benefits. We propose a Contract-Based model of moral cognition that regards artificial systems as relational systems that are subject to a social contract. To articulate this social contract, we draw from contractualism, an impartial ethical framework that evaluates the appropriateness of behaviors based on whether they can be justified to others. In its current form, the Contract-based model characterizes artificial systems as moral agents bound to obligations towards humans. Specifically, this model allows artificial systems to make moral evaluations by estimating the relevance each affected individual assigns to the norms transgressed by an action. It can also learn from human feedback, which is used to generate new norms and update the relevance of different norms in different social groups and types of relationships. The model’s ability to justify their choices to humans, together with the central role of human feedback in moral evaluation and learning, makes this model suitable for supporting human autonomy and fairness in human-to-robot interactions. As human relationships with artificial agents evolve, the Contract-Based model could also incorporate new terms in the social contract between humans and machines, including terms that confer artificial agents a status as moral patients.

The preprint is here.

Here are some thoughts:

The article proposes a Contract-Based model of moral cognition for artificial intelligence (AI) systems, drawing from the ethical framework of contractualism, which evaluates actions based on their justifiability to others. This model views AI systems as relational entities bound by a social contract with humans, allowing them to make moral evaluations by estimating the relevance of norms to affected individuals and learning from human feedback to generate and update these norms. The model is designed to support human autonomy and fairness in human-robot interactions and can also function as moral enhancers to assist humans in moral decision-making in human-human interactions. However, the use of moral enhancers raises ethical concerns about autonomy, responsibility, and potential unintended consequences. Additionally, the article suggests that as human relationships with AI evolve, the model could incorporate new terms in the social contract, potentially recognizing AI systems as moral patients. This forward-thinking approach anticipates the complex ethical questions that may arise as AI becomes more integrated into daily life.

Explainable AI lacks regulative reasons: why AI and human decision-making are not equally opaque

Peters, U.

AI Ethics 3, 963–974 (2023).

https://doi.org/10.1007/s43681-022-00217-w

Abstract

Many artificial intelligence (AI) systems currently used for decision-making are opaque, i.e., the internal factors that determine their decisions are not fully known to people due to the systems’ computational complexity. In response to this problem, several researchers have argued that human decision-making is equally opaque and since simplifying, reason-giving explanations (rather than exhaustive causal accounts) of a decision are typically viewed as sufficient in the human case, the same should hold for algorithmic decision-making. Here, I contend that this argument overlooks that human decision-making is sometimes significantly more transparent and trustworthy than algorithmic decision-making. This is because when people explain their decisions by giving reasons for them, this frequently prompts those giving the reasons to govern or regulate themselves so as to think and act in ways that confirm their reason reports. AI explanation systems lack this self-regulative feature. Overlooking it when comparing algorithmic and human decision-making can result in underestimations of the transparency of human decision-making and in the development of explainable AI that may mislead people by activating generally warranted beliefs about the regulative dimension of reason-giving.

The article is linked above.

Here are some thoughts:

This article delves into the ethics of transparency in algorithmic decision-making (ADM) versus human decision-making (HDM). While both can be opaque, the author argues HDM offers more trustworthiness due to "mindshaping." This theory suggests that explaining decisions, even if the thought process is unclear, influences future human behavior to align with the explanation. This self-regulation is absent in AI, potentially making opaque ADM less trustworthy. The text emphasizes that explanations serve a purpose beyond just understanding the process. It raises concerns about "deceptive AI" with misleading explanations and warns against underestimating the transparency of HDM due to its inherent ability to explain. Key ethical considerations include the need for further research on mindshaping's impact on bias and the limitations of explanations in both ADM and HDM.  Ultimately, the passage highlights the importance of developing explainable AI that goes beyond mere justification, while also emphasizing fairness, accountability, and responsible use of explanations in building trustworthy AI systems.

Is artificial consciousness achievable? Lessons from the human brain

Farisco, M., Evers, K., & Changeux, J.

(2024, April 18). arXiv.org.

https://arxiv.org/abs/2405.04540

Abstract

We here analyse the question of developing artificial consciousness from an evolutionary perspective, taking the evolution of the human brain and its relation with consciousness as a reference model. This kind of analysis reveals several structural and functional features of the human brain that appear to be key for reaching human-like complex conscious experience and that current research on Artificial Intelligence (AI) should take into account in its attempt to develop systems capable of conscious processing. We argue that, even if AI is limited in its ability to emulate human consciousness for both intrinsic (structural and architectural) and extrinsic (related to the current stage of scientific and technological knowledge) reasons, taking inspiration from those characteristics of the brain that make conscious processing possible and/or modulate it, is a potentially promising strategy towards developing conscious AI. Also, it is theoretically possible that AI research can develop partial or potentially alternative forms of consciousness that is qualitatively different from the human, and that may be either more or less sophisticated depending on the perspectives. Therefore, we recommend neuroscience-inspired caution in talking about artificial consciousness: since the use of the same word consciousness for humans and AI becomes ambiguous and potentially misleading, we propose to clearly specify what is common and what differs in AI conscious processing from full human conscious experience.

The research is here.

Here are some thoughts:

Biological Basis of Consciousness

The text emphasizes the critical role of the brain's neurobiological complexity, including molecular diversity, neuronal morphology, neurotransmitters, and connectivity patterns, in enabling conscious experience. It argues that these intricate biological mechanisms are often overlooked or oversimplified in current artificial intelligence (AI) and computational models.

Developmental and Evolutionary Aspects

The text highlights the importance of epigenetic brain development and the "non-genetic evolution" that shapes the brain's connectivity through experience-dependent synaptic selection and pruning processes. This variability and incorporation of individual experiences into the brain's "hardware" is seen as a key aspect of human consciousness that is challenging to replicate in artificial systems.

Cultural Transmission and Acquisition

The text discusses how cultural abilities, such as reading and writing, are epigenetically acquired and transmitted across generations, shaping the brain's functional connectomics. This spontaneous genesis and creative re-elaboration of culture is presented as a distinct feature of human consciousness that current AI systems struggle to capture fully.

Limitations of Current AI Approaches

While acknowledging the impressive capabilities of recent AI developments like large language models (LLMs), the text argues that they fundamentally differ from human cognition and conscious processing. It suggests that AI systems may emulate certain aspects of language and reasoning through statistical patterns and parallel processing, but lack the semantic understanding, meaning attribution, and creative re-elaboration that characterize human consciousness.

Beliefs versus Reality: People Overestimate the Actual Dishonesty of Others

Martuza, J., Thorbjørnsen, H., & Sjåstad, H.

(2024, June 27).

https://doi.org/10.31234/osf.io/nm2cz

Abstract

Beliefs about the dishonesty of other people can shape our collective behavior in powerful ways. How accurate are these beliefs? And do we believe that other people are similarly, more, or less dishonest than they truly are? In a research program on moral decision-making consisting of 31 different effects (N = 8,127), initially designed to test a broad collection of different hypotheses, participants were placed in various situations where they could lie for personal gain, without any repercussions or detection risk. Crucially, we also asked all participants to estimate what percentage of other people would lie in a similar situation. Conducting an internal meta-analysis across all experiments, including both incentivized choice experiments and hypothetical marketplace scenarios, the results revealed a substantial overestimation of others' dishonesty by an average of 14 percentage points (meta-analytic effect: Hedge’s g = 0.58). That is, people are less dishonest than we tend to think, in which about 70% of our participants told the truth despite having a financial incentive to lie. These findings reveal a pervasive tendency to overestimate the actual dishonesty of other people, suggesting a biased belief that the world is less moral than it truly is. Does belief accuracy matter? In a new experiment using an information treatment based on our meta-analytic results (N = 981), providing correct information about actual honesty levels did not only lead to more positive honesty beliefs, but also to a broader enhancement of pro-social expectations – including higher interpersonal trust, fairness and helpfulness beliefs, and reduced cynicism.

My interpretation:

The meta-analysis reveals a significant overestimation of others' dishonesty by an average of 14 percentage points, with a meta-analytic effect of Hedge's g = 0.58. This finding, based on 31 effect sizes and over 8,000 participants, supports the overestimation hypothesis of dishonesty and proves robust across variations in methods and procedures.

Further research demonstrates the impact of correcting these biased beliefs. Informing people about actual honesty rates led to more positive pro-social expectations, including higher interpersonal trust, increased fairness and helpfulness beliefs, and lower cynicism. These findings have important implications, challenging the 'wisdom of crowds' approach in moral judgment and contributing to our understanding of the mixed effects of social norm nudges.

The study highlights the importance of data-based views for leaders and policymakers. However, it's crucial to note the limitations of this research, including its focus on American participants, the potential gap between experimental tasks and real-world complexity, and the use of mostly one-shot belief reports.

In conclusion, this research provides robust evidence that people tend to overestimate others' dishonesty. Moreover, correcting these biased beliefs can positively impact social thinking and expectations, potentially leading to more trusting and cooperative societies.  Important points for clinical psychologists.

How perceived polarization predicts attitude moralization (and vice versa): A four-wave longitudinal study during the 2020 U.S. election

D'Amore, C., van Zomeren, M., & Koudenburg, N.

(2024). Journal of personality and social psychology,

126(4), 624–642.

https://doi.org/10.1037/pspi0000454

Abstract

Within structurally polarized and dynamic contexts, such as the U.S. 2020 presidential elections, the moralization of individuals' attitudes on a specific topic (e.g., climate policy) can dangerously escalate disagreements between groups into zero-sum conflict. However, limited knowledge exists regarding the factors that influence individuals' tendency to moralize their attitudes over time, and what the role of structural polarization is in this psychological process. Our objective is to test a theoretically integrative model of when and how perceived polarization is related to attitude moralization over time within the polarized context of the U.S. 2020 presidential elections and explore reciprocal feedback loops to understand the dynamic relationship between polarization and moralization over time. Our model predicts that, when repeatedly faced with outgroup expressions in the news, individuals' perceptions of polarization will predict within-person attitude moralization over time via strengthening their value-protective responses to these expressions (i.e., perceiving dyadic harm and experiencing negative moral emotions toward the outgroup). To test our model, we conducted a four-wave, 4-month longitudinal study among Biden supporters (N = 1,236) and Trump supporters (N = 617). The results of the within-person analyses generally supported the model's hypotheses across both samples and various attitude topics. Furthermore, cross-lagged structural equation models explored reciprocal influences, revealing positive feedback loops between structural polarization and attitude moralization over time. Our findings thus indicate that perceived polarization strengthens attitude moralization (and vice versa) over time-a dynamic process that helps to explain how nonmoralized conflict between groups can evolve into zero-sum conflict during periods of intense polarization.

Here is my interpretation of the study:

This research on the U.S. 2020 election campaign explored how people strengthen their moral beliefs over time in a polarized environment. They found that as individuals perceive more polarization and political homogeneity in their networks, they are more likely to develop strong moral convictions, especially when they see harm from the opposing political group and feel negative emotions towards them.

The study showed that increased perceptions of polarization lead to stronger moralization of attitudes across various topics. This process is reinforced over time by a feedback loop where perceived polarization and moralization influence each other. These findings were consistent among both Biden and Trump supporters.

In summary, the study confirms that polarization plays a crucial role in shaping strong moral attitudes, driving political engagement and conflict. This research highlights the need to understand the social factors that turn political disagreements into moral conflicts and suggests areas for future exploration.

Predicting Radicalism After Perceived Injustice: The Role of Separatist Identity, Sacred Values, and Police Violence

Pretus, C., Sheikh, H., Hamid, N., & Atran, S. (2023).

Journal of Social and Political Psychology, 

11(2), 730–746. https://doi.org/10.5964/jspp.11255

Abstract

Perceptions of injustice are central to fueling violent political action, though not everyone within a social movement will support violence in response to collective grievances. So who supports violence and who doesn't after perceived injustice? To address this question, we followed up on the same individuals (N = 805) before and after a court decision in Catalonia (Spain) sentencing nine separatist leaders to prison, an event that led to mass violent and nonviolent protests. We tested three hypotheses by combining classical theories of collective action and more recent extremism models and found support for all three hypotheses. Namely, individuals who exhibited steeper increases in radicalism (controlling for activism) after the court ruling were those who had previously experienced police violence (social dynamics hypothesis), those who identified as separatists (separatist identity hypothesis), and those who held Catalan 2 independence as a sacred value (sacred value hypothesis). Our findings offer a complex picture of real-world conflict settings, where the three evaluated factors seem to be intertwined. We discuss potential venues to restore inter-group relations after perceived injustice, with an assessment of how likely these strategies are to succeed based on the three adopted perspectives.

Here are some thoughts:

This study investigated what makes people who feel wronged turn to violence. Researchers studied people in Catalonia, Spain, following a court ruling that many Catalans felt was unfair. They found that people who felt more strongly about Catalan independence (separatist identity and sacred value) and had experienced police violence in the past were more likely to report increased support for violence after the court ruling.

The study suggests that these three factors may reinforce each other. For instance, experiencing police violence might strengthen someone's separatist identity and make them see independence as a more sacred value.

The researchers acknowledge limitations like not having a control group that wasn't exposed to the court ruling. They also note their findings may not apply to groups that see non-violence as a core value. Future studies could look at these factors across different cultures and social movements.

Overall, the study suggests that a sense of injustice, combined with a strong group identity and a feeling that the group's values are under attack, can make people more likely to support violence.

Emotions explain diferences in the difusion of true vs. false social media rumors

Pröllochs, N., Bär, D. & Feuerriegel, S.

Sci Rep 11, 22721 (2021).

https://doi.org/10.1038/s41598-021-01813-2

Abstract

False rumors (often termed “fake news”) on social media pose a signifcant threat to modern societies. However, potential reasons for the widespread difusion of false rumors have been underexplored. In this work, we analyze whether sentiment words, as well as diferent emotional words, in social media content explain diferences in the spread of true vs. false rumors. For this purpose, we collected N = 126,301 rumor cascades from Twitter, comprising more than 4.5 million retweets that have been fact-checked for veracity. We then categorized the language in social media content to (1) sentiment (i.e., positive vs. negative) and (2) eight basic emotions (i. e., anger, anticipation, disgust, fear, joy, trust, sadness, and surprise). We find that sentiment and basic emotions explain differences in the structural properties of true vs. false rumor cascades. False rumors (as compared to true rumors) are more likely to go viral if they convey a higher proportion of terms associated with a positive sentiment.  Further, false rumors are viral when embedding emotional words classifed as trust, anticipation, or anger. All else being equal, false rumors conveying one standard deviation more positive sentiment have a 37.58% longer lifetime and reach 61.44% more users. Our fndings ofer insights into how true vs. false rumors spread and highlight the importance of managing emotions in social media content.

Here are some thoughts:

This research analyzes how language used in social media posts influences the spread of rumors, specifically true vs. false ones. The study focuses on sentiment (positive vs. negative) and basic emotions (anger, anticipation, trust) in online content. They found that positive language and emotions like anger, anticipation, and trust are linked to a wider spread of false rumors. This is because emotions are known to influence how online content is shared.

Final thought: When we encounter a rumor that evokes emotions like anger, anticipation, or trust, it is more likely to stand out to us. This increased salience makes us more likely to share the rumor, even if it is false.