arXiv subjects

Tiago Domingos

Publications and source records attributed to Tiago Domingos.

4 recordsLinked to original sources

arXiv2026

The impacts of artificial intelligence on environmental sustainability and human well-being

Artificial intelligence (AI) is increasingly being described as a transformative general-purpose technology, yet its impacts on environmental sustainability and human well-being remain poorly understood. Here, we conduct a systematic review of 1,291 studies selected from 6,655 records to map how the literature assesses these impacts. We find that current research provides a fragmented and uneven account of AI's consequences for human well-being and the environment. Environmental studies focus narrowly on energy use and CO2 emissions (72%) and rarely consider systemic effects (11%), while well-being studies are predominantly conceptual and overlook subjective well-being, cognitive capabilities, and upstream supply-chain impacts. Strikingly, 82% of environmental studies portray AI's impacts as positive, but this finding is driven by the large number of application-level studies. Well-being analyses show a near-even split (44% positive; 46% negative). However, this split masks differences across well-being dimensions: impacts on income and health are generally expected to be positive, whereas impacts on inequality, social cohesion, and employment are expected to be negative. Based on our findings, we identify important priorities for future research: environmental assessments should consider systemic effects and indicators beyond energy and CO2, while well-being research should prioritise empirical analysis. More fundamentally, research must consider environmental sustainability and human well-being together, recognising that AI's impacts are deeply interconnected. Towards this aim, we extend an existing three-level framework for the climate impacts of AI to include broader environmental impacts and human well-being, providing a unified basis to assess AI impacts and guide its development towards environmental sustainability and human flourishing.

cs.CY↗

arXiv2023

DeepThought: An Architecture for Autonomous Self-motivated Systems

The ability of large language models (LLMs) to engage in credible dialogues with humans, taking into account the training data and the context of the conversation, has raised discussions about their ability to exhibit intrinsic motivations, agency, or even some degree of consciousness. We argue that the internal architecture of LLMs and their finite and volatile state cannot support any of these properties. By combining insights from complementary learning systems, global neuronal workspace, and attention schema theories, we propose to integrate LLMs and other deep learning systems into an architecture for cognitive language agents able to exhibit properties akin to agency, self-motivation, even some features of meta-cognition.

cs.AI↗

arXiv2005

Harvesting in a resource dependent age structured Leslie type population model

We analyse the effect of harvesting in a resource dependent age structured population model, deriving the conditions for the existence of a stable steady state as a function of fertility coefficients, harvesting mortality and carrying capacity of the resources. Under the effect of proportional harvest, we give a sufficient condition for a population to extinguish, and we show that the magnitude of proportional harvest depends on the resources available to the population. We show that the harvesting yield can be periodic, quasi-periodic or chaotic, depending on the dynamics of the harvested population. For populations with large fertility numbers, small harvesting mortality leads to abrupt extinction, but larger harvesting mortality leads to controlled population numbers by avoiding over consumption of resources. Harvesting can be a strategy in order to stabilise periodic or quasi-periodic oscillations in the number of individuals of a population.

q-bio.PE↗

arXiv1999

A General Approach to the Modelling of Trophic Chains

Based on the law of mass action (and its microscopic foundation) and mass conservation, we present here a method to derive consistent dynamic models for the time evolution of systems with an arbitrary number of species. Equations are derived through a mechanistic description, ensuring that all parameters have ecological meaning. After discussing the biological mechanisms associated to the logistic and Lotka-Volterra equations, we show how to derive general models for trophic chains, including the effects of internal states at fast time scales. We show that conformity with the mass action law leads to different functional forms for the Lotka-Volterra and trophic chain models. We use mass conservation to recover the concept of carrying capacity for an arbitrary food chain.

adap-org↗