This paper argues that large language models have a valuable scientific role to play in serving as scientific models of public languages. Linguistic study should not only be concerned with the cognitive processes behind linguistic competence, but also with language understood as an external, social entity. Once this is recognized, the value of large language models as scientific models becomes clear. This paper defends the position against a number of arguments to the effect that language models (...) provide no linguistic insight. Building upon Weisberg’s (Brit J Philos Sci 58(2):207–233, 2007) notion of a model construal, it is then argued that recent work in computational linguistics to better understand the inner workings of large language models can be used to develop a model construal for large language models as models of a language. (shrink)

Is language understanding a special case of social cognition? To help evaluate this view, we can formalize it as the rational speech-act theory: Listeners assume that speakers choose their utterances approximately optimally, and listeners interpret an utterance by using Bayesian inference to “invert” this model of the speaker. We apply this framework to model scalar implicature (“some” implies “not all,” and “N” implies “not more than N”). This model predicts an interaction between the speaker's knowledge state and the listener's (...) interpretation. We test these predictions in two experiments and find good fit between model predictions and human judgments. (shrink)

Mental model construction is supposed to be a useful cognitive devise for learning. Beyond human capacity of constructing mental models, scientists construct complex explanations about phenomena, named scientific or theoretical models. In this work we revisit three vissions: the first one concern about the polisemic term “model”. Our proposal is to discriminate between “mental models” and “explicit models”, being the former those “imaginistic” ideas constructed in scientists’—o teachers—minds, and the latter those teaching devices expressed in different languages that tend to (...) communicate any “scientific model”. From this point of view, the class is considered a place where teachers’ mental models should be learned by novice students by decoding their teaching devices which are expressed in different languages. Other proposal of this work claims to distinguish the term “representation” with respect to its artistical or instrumental origin, highlighting that they are types of teaching devices and that artistical representations are always analogies. Finally, data about the construction of freshmen’s wrong mental models related to the use of the analogy between the chemical combustion and the global process of cellular respiration from glucose is presented to reinforce previous epistemological reflections. (shrink)

As part of a larger research study, we present model language for disclosing financial interests in clinical research to potential research participants, and we describe the empirical basis and theoretical assumptions used in developing the language. The empirical process for creating appropriate disclosure language resulted in a generic disclosure statement for cases in which no risk to participants’ welfare or the scientific integrity of the research is expected, and nine more specific disclosure statements for cases in which (...) some risk is expected. The disclosure statements are not meant to be canonical, but were instead designed to reflect the typical situations in which disclosure of financial interest might be considered by an institutional review board or conflict of interest committee. Individual institutions could modify key phrases to suit their purposes, and others could use the language in future empirical work on informed consent to better refine the options for disclosing financial interests in clinical research. (shrink)

This paper argues that large language models have a valuable scientific role to play in serving as scientific models of public languages. Linguistic study should not only be concerned with the cognitive processes behind linguistic competence, but also with language understood as an external, social entity. Once this is recognized, the value of large language models as scientific models becomes clear. This paper defends the position against a number of arguments to the effect that language models (...) provide no linguistic insight. Building upon Weisberg’s (2007) notion of a model construal, it is then argued that recent work in computational linguistics to better understand the inner workings of large language models can be used to develop a model construal for large language models as models of a language. (shrink)

Conceptual modeling languages are purposeful artifacts, hence their design should also start from the purpose that they serve. Such purposeful design addresses the requirements engineering...

In his Memoir of Wittgenstein Professor Malcolm describes the occasion on which, as far as he knows, the idea that as an activity language is a game, or that ‘games are played with words’, first occurred to Wittgenstein. Wittgenstein was passing a playing field where there was a game of football in progress. As he watched the game, the thought suddenly flashed into his mind, ‘We play games with words !’ This account may be compared with that given by (...) Professor von Wright, on the basis of Wittgenstein's own report, of the way in which he was with equal suddenness struck with the thought that a proposition is a ‘picture’. He was looking at a diagram which showed the positions of certain persons and vehicles involved in an accident—‘ That is a proposition!’ was the thought which on this earlier occasion darted into Wittgenstein's mind. Thus, it seems, was germinated the theory developed in the Tractatus that propositions, that is to say, meaningful propositions, are ‘pictures of reality’. (shrink)

In quantum information and computation, entanglement is a resource. When combining concepts, the application of entanglement outside of micro-physical systems is an useful tool. We suggest new cognitive image-based tests that do not need to be translated. No prior knowledge of terms related to the concepts is required, therefore the choice is more intuitive. We examine the merging of two concepts that establish non-classical statistical correlation and present an entanglement-aware vector encoding algorithm. This research’s added value results in an automated (...) system that teaches artificial intelligence to identify and handle entangled concepts. (shrink)

An essential aspect of conceptual data modeling methodologies is the language’s expressiveness so as to represent the subject domain as precise as possible to obtain good quality models and, consequently, software. To gain better insight in the characteristics of the main conceptual modeling languages, we conducted a comparison between ORM, ORM2, UML, ER, and EER with the aid of Description Logic languages of the DLR family and the new formally defined generic conceptual data modeling language CMcom that is (...) based on DLRifd. ORM, ER, EER, and UML class diagrams are proper fragments of ORM2 and CMcom has the most expressive common denominator with these languages. CMcom simplifies prospects for automated, online, interoperability among the considered languages so that modelers not only can continue using their preferred modeling language yet be compatible with the other ones, but also have a common ground that eases database and software integration based on commonly used conceptual data models. (shrink)

From the side of modelers and early-adopter industry, interest in reasoning over conceptual models and other online usage of conceptual models is growing. To obtain a more precise insight in the characteristics of the main conceptual modeling languages, we define the (semi-)standardized ORM, ORM2, UML, ER, and EER diagram languages in terms of the new generic conceptual data modeling language CMcom that is based on the DL language DLRifd. CMcom has the most expressive common denominator with these languages. (...) CMcom advances prospects for automated, online, interoperability among diverse conceptual data models and ensures compatibility with and between industrygrade conceptual data modeling languages. (shrink)

Bilingual dictionaries are an important aid in foreign language acquisition and in interlingual communication. However, when speaking and writing one needs to be able to formulate whole sentences instead of using single words. The term syntagmatics encompasses all that surrounds a word in a sentence. The monograph Syntagmatics in the Bilingual Dictionary explores what one needs to know about a word to use it correctly in a sentence and how a dictionary should be structured to convey this information.

LLMs are, in the first instance, models of the statistical distribution of tokens in the vast linguistic corpus they have been trained on. But their often surprising emergent capabilities raise the question of how much understanding of the extralinguistic world LLMs can glean from this statistical distribution of words alone. Here, I explore and evaluate the idea that the probability distribution of words in the public corpus offers a window onto the conditional structure of the world. To become a good (...) next-token predictor, an LLM has to become a good pattern completer, and the patterns laid up in language mirror the patterns embodied in the world to a considerable extent. When this is the case—and this is no empty condition—it allows LLMs to compress into their weights two distinct, but complementary forms of understanding of the world. The first form of understanding stored is an understanding of which properties tend to cluster together; I spell out this idea by drawing on Millikan’s account of a multidimensional clumpy world. The second is an understanding of the inferential structure that propositions about these property clusters tend to be enmeshed in; I spell out this idea by drawing on Brandom’s account of the isomorphism between deontic normative conceptual relations of incompatibility and consequence among commitments and alethic modal relations of incompatibility and consequence among states of affairs. On the resulting picture, word models can be world models insofar as linguistic patterns track real patterns between facts. (shrink)

Aristotle was probably one of the first who noted the complex structure of language in his treatise «Categories», stating that what appears simple is in fact the result of brainwork. However, based on the result alone, it is impossible to reconstruct the entire complex structure of language production. Assuming words as units of meaning, it is impossible to reconstruct meaning solely by manipulating words in a subjectless space. Therefore, modeling language requires constructive and substrative foundations of a (...) different quality than those found in contemporary machines. (shrink)

Do large language models like Chat-GPT or Claude meaningfully use the words they produce? Or are they merely clever prediction machines, simulating language use by producing statistically plausible text? There have already been some initial attempts to answer this question by showing that these models meet the criteria for entering meaningful states according to metasemantic theories of mental content. In this paper, I will argue for a different approach—that we should instead consider whether language models meet the (...) criteria given by our best metasemantic theories of linguistic content. In that vein, I will illustrate how this can be done by applying two such theories to the case of language models: Gareth Evans’ ( 1982 ) account of naming practices and Ruth Millikan’s ( 1984, 2004, 2005 ) teleosemantics. In doing so, I will argue that it is a mistake to think that the failure of LLMs to meet plausible conditions for mental intentionality thereby renders their outputs meaningless, and that a distinguishing feature of linguistic intentionality—dependency on a pre-existing linguistic system—allows for the plausible result that LLM outputs are meaningful. (shrink)

To what degree can language be acquired from linguistic input alone? This question has vexed scholars for millennia and is still a major focus of debate in the cognitive science of language. The complexity of human language has hampered progress because studies of language–especially those involving computational modeling–have only been able to deal with small fragments of our linguistic skills. We suggest that the most recent generation of Large Language Models (LLMs) might finally provide the (...) computational tools to determine empirically how much of the human language ability can be acquired from linguistic experience. LLMs are sophisticated deep learning architectures trained on vast amounts of natural language data, enabling them to perform an impressive range of linguistic tasks. We argue that, despite their clear semantic and pragmatic limitations, LLMs have already demonstrated that human-like grammatical language can be acquired without the need for a built-in grammar. Thus, while there is still much to learn about how humans acquire and use language, LLMs provide full-fledged computational models for cognitive scientists to empirically evaluate just how far statistical learning might take us in explaining the full complexity of human language. (shrink)

Large Language Models (LLMs) have been transformative. They are pre-trained foundational models that are self-supervised and can be adapted with fine tuning to a wide range of natural language tasks, each of which previously would have required a separate network model. This is one step closer to the extraordinary versatility of human language. GPT-3 and more recently LaMDA can carry on dialogs with humans on many topics after minimal priming with a few examples. However, there has been (...) a wide range of reactions and debate on whether these LLMs understand what they are saying or exhibit signs of intelligence. This high variance is exhibited in three interviews with LLMs reaching wildly different conclusions. A new possibility was uncovered that could explain this divergence. What appears to be intelligence in LLMs may in fact be a mirror that reflects the intelligence of the interviewer, a remarkable twist that could be considered a Reverse Turing Test. If so, then by studying interviews we may be learning more about the intelligence and beliefs of the interviewer than the intelligence of the LLMs. As LLMs become more capable they may transform the way we interact with machines and how they interact with each other. Increasingly, LLMs are being coupled with sensorimotor devices. LLMs can talk the talk, but can they walk the walk? A road map for achieving artificial general autonomy is outlined with seven major improvements inspired by brain systems. LLMs could be used to uncover new insights into brain function by downloading brain data during natural behaviors. (shrink)

This paper defends the thesis that LLM hallucinations are best explained as a truth representation problem: Current models lack an internal representation of propositions as truth-bearers, so truth and falsity cannot constrain generation in the way factual discourse requires. It begins by surveying leading explanations—computational limits on self-verification, deficiencies in training data as truth sources, and architectural factors—and argues that they converge on the same underlying representational deficit. Next, it reconstructs the philosophical background of current LLM design, showing how optimization (...) for fluent continuation aligns with coherence-style evaluation and with a broadly structuralist, relational semantics, before turning to David Chalmers’s recent attempt to secure propositional interpretability by drawing on Davidson/Lewis-style radical interpretation and by locating propositional content in “middle-layer” structures; it argues that this approach downplays the ubiquity of hallucination and inherits instability from post-training edits. Finally, the paper offers a positive proposal: Atomic propositions should be represented in the basic vector layer, reviving a logical-atomist program as a principled route to reducing hallucination. (shrink)

The potential for Large Language Models (LLMs) to attain technological singularity—the point at which artificial intelligence (AI) surpasses human intellect and autonomously improves itself—is a critical concern in AI research. This paper explores the feasibility of current LLMs achieving singularity by examining the philosophical and practical requirements for such a development. We begin with a historical overview of AI and intelligence amplification, tracing the evolution of LLMs from their origins to state-of-the-art models. We then proposes a theoretical framework to (...) assess whether existing LLM technologies could satisfy the conditions for singularity, with a focus on Recursive Self-Improvement (RSI) and autonomous code generation. We integrate key component technologies, such as Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO), into our analysis, illustrating how these could enable LLMs to independently enhance their reasoning and problem-solving capabilities. By mapping out a potential singularity model lifecycle and examining the dynamics of exponential growth models, we elucidate the conditions under which LLMs might self-replicate and rapidly escalate their intelligence. We conclude with a discussion of the ethical and safety implications of such developments, underscoring the need for responsible and controlled advancement in AI research to mitigate existential risks. Our work aims to contribute to the ongoing dialogue on the future of AI and the critical importance of proactive measures to ensure its beneficial development. (shrink)

Are LLMs cultural technologies like photocopiers or printing presses, which transmit information but cannot create new content? A challenge for this idea, which we call bibliotechnism, is that LLMs generate novel text. We begin with a defense of bibliotechnism, showing how even novel text may inherit its meaning from original human-generated text. We then argue that bibliotechnism faces an independent challenge from examples in which LLMs generate novel reference, using new names to refer to new entities. Such examples could be (...) explained if LLMs were not cultural technologies but had beliefs, desires, and intentions. According to interpretationism in the philosophy of mind, a system has such attitudes if and only if its behavior is well explained by the hypothesis that it does. Interpretationists may hold that LLMs have attitudes, and thus have a simple solution to the novel reference problem. We emphasize, however, that interpretationism is compatible with very simple creatures having attitudes and differs sharply from views that presuppose these attitudes require onsciousness, sentience, or intelligence (topics about which we make no claims). (shrink)

We discuss potential biorisks from large language models (LLMs). AI assistants based on LLMs such as ChatGPT have been shown to significantly reduce barriers to entry for actors wishing to synthesize dangerous, potentially novel pathogens and chemical weapons. The harms from deploying such bioagents could be further magnified by AI-assisted misinformation. We endorse several policy responses to these dangers, including prerelease evaluations of biomedical AIs by subject-matter experts, enhanced surveillance and lab screening procedures, restrictions on AI training data, and (...) access controls on biomedical AI. We conclude with a suggestion about future research directions in bioethics. (shrink)

To model behavioral and neural correlates of language comprehension in naturalistic environments, researchers have turned to broad‐coverage tools from natural‐language processing and machine learning. Where syntactic structure is explicitly modeled, prior work has relied predominantly on context‐free grammars (CFGs), yet such formalisms are not sufficiently expressive for human languages. Combinatory categorial grammars (CCGs) are sufficiently expressive directly compositional models of grammar with flexible constituency that affords incremental interpretation. In this work, we evaluate whether a more expressive CCG provides (...) a better model than a CFG for human neural signals collected with functional magnetic resonance imaging (fMRI) while participants listen to an audiobook story. We further test between variants of CCG that differ in how they handle optional adjuncts. These evaluations are carried out against a baseline that includes estimates of next‐word predictability from a transformer neural network language model. Such a comparison reveals unique contributions of CCG structure‐building predominantly in the left posterior temporal lobe: CCG‐derived measures offer a superior fit to neural signals compared to those derived from a CFG. These effects are spatially distinct from bilateral superior temporal effects that are unique to predictability. Neural effects for structure‐building are thus separable from predictability during naturalistic listening, and those effects are best characterized by a grammar whose expressive power is motivated on independent linguistic grounds. (shrink)

Gender studies has generated numerous questions around “neutral” forms, such as the concept of “Self”. The aim of this analysis is to highlight how “neutral” forms are central to the reiteration of the binary model and the dominance of “man2”. Historically, man is the archetype, placing his supremacy as part of the natural order of things. Inserted into this model, many thinkers have considered the male as the transcendental gender, so, elevating the masculine as universal, a-sexed and decorporealised. In this (...) way, man has convinced himself that he’s not conditioned by his masculinity and can speak for all humanity, becoming the logosthrough which he declines the rest. Man, therefore, has made himself “neutral” – both in the conceptual sense and in the grammatical structure (particularly of binary languages such as Italian) – by taking control of language. Through it, he orders and constitutes the world, developing dichotomies and signifying anatomical bodies. In this research, it is intended to work on language and the relationship between “neutral” and power, emphasizing how that relationship is central to the reproduction of the patriarchal model. Through a critical philosophical reflection, which sees a logical-linguistic and historical-processual methodological structure, the discourse of the I-neutral will be crossed, pointing out how that “neutral” is in truth male. The concluding goal will be to, through “fluid” visions, build the basis for a language that is truly neutral and ready to embrace multiplicity, without relations of domination; a language really inclusive and not a reiterator of the patriarchal model. (shrink)

Standard models of lexical production assume that speakers access representations of meaning, grammar, and different aspects of sound in a roughly sequential manner (whether or not they admit cascading or interactivity). In contrast, we review evidence for a parallel activation model in which these representations are accessed in parallel. According to this account, word learning involves the binding of the meaning, grammar, and sound of a word into a single representation. This representation is then activated as a whole during production, (...) and so all linguistic components are available simultaneously. We then note that language comprehension involves extensive use of prediction and argue that comprehenders use production mechanisms to determine (roughly) what they would say next if they were speaking. So far, theories of prediction-by-production have assumed sequential lexical production. We therefore reinterpret such evidence in terms of parallel lexical production. (shrink)

Current work in language models (LMs) helps us speed up or even skip thinking by accelerating and automating cognitive work. But can LMs help us with critical thinking -- thinking in deeper, more reflective ways which challenge assumptions, clarify ideas, and engineer new concepts? We treat philosophy as a case study in critical thinking, and interview 21 professional philosophers about how they engage in critical thinking and on their experiences with LMs. We find that philosophers do not find LMs (...) to be useful because they lack a sense of selfhood (memory, beliefs, consistency) and initiative (curiosity, proactivity). We propose the selfhood-initiative model for critical thinking tools to characterize this gap. Using the model, we formulate three roles LMs could play as critical thinking tools: the Interlocutor, the Monitor, and the Respondent. We hope that our work inspires LM researchers to further develop LMs as critical thinking tools and philosophers and other 'critical thinkers' to imagine intellectually substantive uses of LMs. (shrink)

This essay explores the possible contribution of Large Language Models (LLMs) to human cognition. It investigates whether human cognition can be enhanced by advanced AI systems such as LLMs. Can LLMs make people as learners smarter, or, on the contrary, make them reason/think less? The author discusses the concepts of human and artificial intelligence and examines LLMs as advanced AI systems, which use deep learning techniques and can be considered as excelling in neural network architectures, data volume, generalisation and (...) scalability. The author suggests that while LLMs can assist in facilitating numerous cognitive tasks, more research and philosophical inquiry is needed to understand whether such kind of AI assistance would make people cultivate human intelligence more, and not less. Presumably, Large Language Models (LLMs) can contribute to human intelligence and cognition just under strict (addressed existing limitations, questioning prompting, time-sensitivity, etc.) conditions. However, it is important that these theoretical considerations could be verified by experimental research. (shrink)

Landgrebe and Smith :2061–2081, 2021) present an unflattering diagnosis of recent advances in what they call language-centric artificial intelligence—perhaps more widely known as natural language processing: The models that are currently employed do not have sufficient expressivity, will not generalize, and are fundamentally unable to induce linguistic semantics, they say. The diagnosis is mainly derived from an analysis of the widely used Transformer architecture. Here I address a number of misunderstandings in their analysis, and present what I take (...) to be a more adequate analysis of the ability of Transformer models to learn natural language semantics. To avoid confusion, I distinguish between inferential and referential semantics. Landgrebe and Smith ’s analysis of the Transformer architecture’s expressivity and generalization concerns inferential semantics. This part of their diagnosis is shown to rely on misunderstandings of technical properties of Transformers. Landgrebe and Smith also claim that referential semantics is unobtainable for Transformer models. In response, I present a non-technical discussion of techniques for grounding Transformer models, giving them referential semantics, even in the absence of supervision. I also present a simple thought experiment to highlight the mechanisms that would lead to referential semantics, and discuss in what sense models that are grounded in this way, can be said to understand language. Finally, I discuss the approach Landgrebe and Smith advocate for, namely manual specification of formal grammars that associate linguistic expressions with logical form. (shrink)

Newly powerful large language models have burst onto the scene, with applications across a wide range of functions. We can now expect to encounter their outputs at rapidly increasing volumes and frequencies. Some commentators claim that large language models are bullshitting, generating convincing output without regard for the truth. If correct, that would make large language models distinctively dangerous discourse participants. Bullshitters not only undermine the norm of truthfulness (by saying false things) but the normative status of (...) truth itself (by treating it as entirely irrelevant). So, do large language models really bullshit? I argue that they can, in the sense of issuing propositional content in response to fact-seeking prompts, without having first assessed that content for truth or falsity. However, I further argue that they need not bullshit, given appropriate guardrails. So, just as with human speakers, the propensity for a large language model to bullshit depends on its own particular make-up. (shrink)

In some recent work, some psychologists, linguists, AI researchers, and philosophers (e.g., Shanahan, 2022; Bender & Koller, 2020; Montemayor, 2021; Bender et al., 2021) have argued that, despite producing convincing human-like linguistic output, large language models do not possess linguistic competence on the ground that they lack communicative intention. Among the proponents of this position, the notion of communicative intention is entertained as the liveliest candidate for a distinguishing characteristic of human cognition vis-a-vis LLMs and as such as at (...) least one of the features which accounts for the characteristic mentality of humans. I demonstrate that such a position is largely driven by strong Gricean assumptions about the nature of linguistic competence that I argue is now known to be empirically untenable. Moreover, I show that once we give up these assumptions and fairly assess the architecture of both LLM and human cognitive architectures, we will find no good reasons to suppose large language models lack communicative intentions. I thus reject communicative intentions as a viable ground on which to understand the difference between LLMs and humans and thus as a characteristic mark of human mentality. (shrink)

Large language models (LLMs) have now entered the realm of medical ethics. In a recent study, Balaset alexamined the performance of GPT-4, a commercially available LLM, assessing its performance in generating responses to diverse medical ethics cases. Their findings reveal that GPT-4 demonstrates an ability to identify and articulate complex medical ethical issues, although its proficiency in encoding the depth of real-world ethical dilemmas remains an avenue for improvement. Investigating the integration of LLMs into medical ethics decision-making appears to (...) be an interesting avenue of research. However, despite the promising trajectory of LLM technology in medicine, it is crucial to exercise caution and refrain from attributing their expertise to medical ethics. Our thesis follows an examination of the nature of expertise and the epistemic limitations that affect LLM technology. As a result, we propose two more fitting applications of LLMs in medical ethics: first, as tools for mining electronic health records or scientific literature, thereby supplementing evidence for resolving medical ethics cases, and second, as educational platforms to foster ethical reflection and critical thinking skills among students and residents. The integration of LLMs in medical ethics, while promising, requires careful consideration of their epistemic limitations. Consequently, a well-considered definition of their role in ethically sensitive decision-making is crucial. (shrink)

Guiding the work of most linguists and philosophers of language today is the assumption that language is governed by prescriptive normative rules. Many believe that it is of the essence of thought itself to follow rules, rules of inference determining the intentional contents of our concepts, and that these rules originate as internalized rules of language. However, exactly what it is for there to be such things as normative rules of language remains distressingly unclear. From what (...) source do these norms flow? What sanctions enforce them? What happens, exactly, if you don't follow the rules? How do children learn the rules?Ruth Millikan presents a radicallly different way of viewing the partial regularities that language displays, the norms and conventions of language. The central norms applying to language, like those norms of function and behavior that account for the survival and proliferation of biological traits, are non-evaluative norms. Specific linguistic forms survive and are reproduced together with co-operative hearer responses because, in a critical mass of cases, these patterns of production and response benefit both speakers and hearers. Conformity is needed only often enough to ensure that the co-operative use constituting the norm - the convention - continues to be copied and hence continues to characterize some interactions of some speaker-hearer pairs.What needs to be reproduced for discursive language forms to survive, it turns out, is not specific conceptual roles but only satisfaction conditions coupled to essential elements of hearer responses. An uncompromising rejection of conceptual analysis as a tool in philosophy results. At the same time the distinction between the propositional content and the force of a linguistic utterance comes into very sharp focus, force emerging as essential to the creation of content rather than as something added to content. The distinction between illocutionary and perlocutionary force, the distinction between linguistic meaning and speaker meaning, and the semantics/pragmatics distinction are each illuminated in new and crisper ways. On the model proposed, neither the intentionality of thought nor the intentionality of language is derived from the other. Processes involved in understanding language are not Gricean but more like direct perception of the world as mediated, for example, through the natural signs contained in the structured light that allows vision. There are also startling implications for pragmatics, and for how children learn language. (shrink)

Artificial intelligence chatbots driven by large language models (LLMs) have the potential to increase public science literacy and support scientific research, as they can quickly summarize complex scientific information in accessible terms. However, when summarizing scientific texts, LLMs may omit details that limit the scope of research conclusions, leading to generalizations of results broader than warranted by the original study. We tested 10 prominent LLMs, including ChatGPT-4o, ChatGPT-4.5, DeepSeek, LLaMA 3.3 70B, and Claude 3.7 Sonnet, comparing 4900 LLM-generated summaries (...) to their original scientific texts. Even when explicitly prompted for accuracy, most LLMs produced broader generalizations of scientific results than those in the original texts, with DeepSeek, ChatGPT-4o, and LLaMA 3.3 70B overgeneralizing in 26–73% of cases. In a direct comparison of LLM-generated and human-authored science summaries, LLM summaries were nearly five times more likely to contain broad generalizations (OR = 4.85, 95% CI [3.06, 7.70], p < 0.001). Notably, newer models tended to perform worse in generalization accuracy than earlier ones. Our results indicate a strong bias in many widely used LLMs towards overgeneralizing scientific conclusions, posing a significant risk of large-scale misinterpretations of research findings. We highlight potential mitigation strategies, including lowering LLM temperature settings and benchmarking LLMs for generalization accuracy. (shrink)

Large language models (LLMs) such as ChatGPT or Google’s Bard have shown significant performance on a variety of text-based tasks, such as summarization, translation, and even the generation of new...

Which is better at doing medical ethics: conversational artificial intelligence bots like ChatGPT or tools based on fuzzy cognitive maps? The article compares the performance of chatbots that rely on large language models to that of our own METHAD algorithm. While both tools approach dilemmas in medical ethics through the lens of Beauchamp and Childress’ mid-level principles, ChatGPT and METHAD differ considerably in the format of their inputs and outputs, in their interpretability, and in the kinds of mistakes that (...) they make. An ideal advisory algorithm would combine their characteristics. (shrink)

This article sets in with the question whether current or foreseeable transformer-based large language models (LLMs), such as the ones powering OpenAI’s ChatGPT, could be language users in a way comparable to humans. It answers the question negatively, presenting the following argument. Apart from niche uses, to use language means to act. But LLMs are unable to act because they lack intentions. This, in turn, is because they are the wrong kind of being: agents with intentions need (...) to be autonomous organisms while LLMs are heteronomous mechanisms. To conclude, the article argues, based on structural aspects of transformer-based LLMs, that these LLMs have taken a first step away from mechanistic artificiality to autonomous self-constitution, which means that these models are (slowly) moving into a direction that someday might result in non-human, but equally non-artificial agents, thus subverting the time-honored Kantian distinction between organism and mechanism. (shrink)

Large Language Models (LLMs) could be a useful tool for lawyers. However, empirical research on their effectiveness in conducting legal tasks is scant. We study securities cases involving cryptocurrencies as one of numerous contexts where AI could support the legal process, studying GPT-3.5’s legal reasoning and ChatGPT’s legal drafting capabilities. We examine whether a) GPT-3.5 can accurately determine which laws are potentially being violated from a fact pattern, and b) whether there is a difference in juror decision-making based on (...) complaints written by a lawyer compared to ChatGPT. We feed fact patterns from real-life cases to GPT-3.5 and evaluate its ability to determine correct potential violations from the scenario and exclude spurious violations. Second, we had mock jurors assess complaints written by ChatGPT and lawyers. GPT-3.5’s legal reasoning skills proved weak, though we expect improvement in future models, particularly given the violations it suggested tended to be correct (it merely missed additional, correct violations). ChatGPT performed better at legal drafting, and jurors’ decisions were not statistically significantly associated with the author of the document upon which they based their decisions. Because GPT-3.5 cannot satisfactorily conduct legal reasoning tasks, it would be unlikely to be able to help lawyers in a meaningful way at this stage. However, ChatGPT’s drafting skills (though, perhaps, still inferior to lawyers) could assist lawyers in providing legal services. Our research is the first to systematically study an LLM’s legal drafting and reasoning capabilities in litigation, as well as in securities law and cryptocurrency-related misconduct. (shrink)

Can large language models produce expert‐quality philosophical texts? To investigate this, we fine‐tuned GPT‐3 with the works of philosopher Daniel Dennett. To evaluate the model, we asked the real Dennett 10 philosophical questions and then posed the same questions to the language model, collecting four responses for each question without cherry‐picking. Experts on Dennett's work succeeded at distinguishing the Dennett‐generated and machine‐generated answers above chance but substantially short of our expectations. Philosophy blog readers performed similarly to the experts, (...) while ordinary research participants were near chance distinguishing GPT‐3's responses from those of an “actual human philosopher”. (shrink)

We explore the capabilities of Large Language Models (LLMs) by comparing the way they gather data with the way humans build knowledge. Here we examine how scientific knowledge is made and compare it with LLMs. The argument is structured by reference to two figures, one representing scientific knowledge and the other LLMs. In a 2014 study, scientists explain how they choose to ignore a ‘fringe science’ paper in the domain in the domain of gravitational wave physics: the decisions are (...) made largely as a result of tacit knowledge built up in social discourse, mostly spoken discourse, within closed groups of experts. It is argued that LLMs cannot or do not currently access such discourse, but it is typical of the early formation of scientific knowledge. LLMs ‘understanding’ builds on written literatures and is therefore insecure in the case of the initial stages of knowledge building. We refer to Colin Fraser’s ‘Dumb Monty Hall problem’ where in 2023 ChatGPT failed though a year or so later LLMs were succeeding. We argue that this is not a matter of improvement in LLMs ability to reason but in the change in the body of human written discourse on which they can draw (or changes being put in by humans ‘by hand’). We then invent a new Monty Hall prompt and compare the responses of a panel of LLMs and a panel of humans: they are starkly different but we explain that the previous mechanisms will soon allow the LLMs to align themselves to humans once more. Finally, we look at ‘overshadowing’ where a settled body of discourse becomes so dominant that LLMs fail to respond to small variations in prompts which render the old answers nonsensical. The ‘intelligence’, we argue, is in the humans not the LLMs. (shrink)

This work develops a regulatory ontology of language that departs from both representational and usage-based paradigms. Rather than treating language as a system of meanings, symbols, or socially enforced rules, the text analyzes language as a historically stabilized mechanism of behavioral regulation operating under conditions of uncertainty. -/- The central claim is that language does not originate from symbolic representation or propositional communication, but from pre-symbolic processes of imitation, synchronization, and the stabilization of coordinated action. Semantic (...) structures, grammatical rules, dictionaries, and alphabets are treated as late-stage externalizations of prior regulatory distinctions, rather than as foundational components of language. -/- The framework is articulated through a set of ontological commitments and their necessary consequences, which collectively delineate principled limits of translation, information-theoretic models of communication, and purely text-based approaches to language learning. Particular attention is given to cases where dominant theoretical assumptions fail, including cross-dictionary translation, alphabet mismatch, cryptographic encoding across incompatible systems, the search for extraterrestrial intelligence (SETI), and contemporary large language models (LLMs). -/- The work does not aim to propose a comprehensive linguistic theory or to compete with existing models within their domains of applicability. Instead, it seeks to expose the ontological boundaries within which such models remain meaningful. Language, on this account, is observable only through the externalized residue of historically viable regulation; beyond this residue lie not semantic gaps or informational noise, but the limits of regulation itself. -/- This text is intended as a foundational contribution for further philosophical, cognitive, and interdisciplinary research, and as a reference point for examining the limits of translation, information, and artificial language competence. (shrink)

As large language models (LLMs) such as ChatGPT, Claude, Gemini, and Perplexity become increasingly ubiquitous as both tools and objects of scientific study, in addition to their established roles as chatbots, text generators and translators, questions about their identity conditions become scientifically as well as philosophically and socially important. This paper is about how to count language models. We argue that much of the emerging literature on these systems presupposes an answer to the question of identity for these (...) AIs but lacks a framework for counting or individuating them. We propose four levels of analysis, inspired by an analogy with software as well as insights from social epistemology. This account produces practical consequences for both research on language models, the ethics of AI, and our everyday talk about them. (shrink)

The National Cancer Institute's (NCI) Thesaurus is a biomedical reference ontology. The NCI Thesaurus is represented using description logic, more specifically Ontylog, a description logic implemen...

A growing chorus of AI researchers and philosophers posit internal representations in large language models (LLMs). But how do these representations relate to the kinds of mental states we routinely ascribe to our fellow humans? While some research has focused on belief- or knowledge- like states in LLMs, there has been comparatively little focus on the question of whether LLMs have intentions. I survey five properties that have been associated with intentions in the philosophical literature, and assess two candidate (...) classes of LLM representations against this set of features. The result is mixed: LLMs have representations that are intention-like in many— perhaps surprising—respects, but they differ from human intentions in important ways. (shrink)

Large language models (LLMs) challenge Chomsky’s long-standing mysterian view of the creative aspect of language use (CALU). By exhibiting fluent, situation-appropriate linguistic behavior and offering concrete mechanistic hypotheses, they provide the first viable scientific models of CALU. We endorse Futrell and Mahowald’s call to integrate LLMs into linguistic inquiry and suggest a bolder aim: elucidating the mechanisms underlying linguistic creativity.

In a recent letter, Dillion et. al (2023) make various suggestions regarding the idea of artificially intelligent systems, such as large language models, replacing human subjects in empirical moral psychology. We argue that human subjects are in various ways indispensable.

[This is an edited version of a keynote talk at the conference on Neural Information Processing Systems (NeurIPS) on November 28, 2022, with some minor additions and subtractions.] There has recently been widespread discussion of whether large language models might be sentient or conscious. Should we take this idea seriously? I will break down the strongest reasons for and against. Given mainstream assumptions in the science of consciousness, there are significant obstacles to consciousness in current models: for example, their (...) lack of recurrent processing, a global workspace, and unified agency. At the same time, it is quite possible that these obstacles will be overcome in the next decade or so. I conclude that while it is somewhat unlikely that current large language models are conscious, we should take seriously the possibility that successors to large language models may be conscious in the not-too-distant future. (shrink)