Web agents, like OpenAI's Operator and Google's Project Mariner, are powerful agentic systems pushing the boundaries of Large Language Models (LLM). They can autonomously interact with the internet at the user's behest, such as navigating websites, filling search masks, and comparing price lists. Though web agent research is thriving, induced sustainability issues remain largely unexplored. To highlight the urgency of this issue, we provide an initial exploration of the energy and CO₂ cost associated with web agents from both a theoretical —via estimation— and an empirical perspective —by benchmarking. Our results show how different philosophies in web agent creation can severely impact the associated expended energy, and that more energy consumed does not necessarily equate to better results. We highlight a lack of transparency regarding disclosing model parameters and processes used for some web agents as a limiting factor when estimating energy consumption. Our work contributes towards a change in thinking of how we evaluate web agents, advocating for dedicated metrics measuring energy consumption in benchmarks.

The increased presence of large language models (LLMs) in educational settings has ignited debates concerning negative repercussions, including overreliance and inadequate task reflection.Our work advocates moderated usage of such models, designed in a way that supports students and encourages critical thinking.We developed two moderated interaction methods with ChatGPT: hintbased assistance and presenting multiple answer choices.In a study with students (N=40) answering physics questions, we compared the effects of our moderated models against two baseline settings: unmoderated ChatGPT access and internet searches.We analyzed the interaction strategies and found that the moderated versions exhibited less unreflected usage (e.g., copy & paste) compared to the unmoderated condition.However, neither ChatGPT-supported condition could match the ratio of reflected usage present in internet searches.Our research highlights the potential benefits of

Monitoring of human activities is an essential capability of many smart systems. In recent years much progress has been achieved. One of the key remaining challenges is the availability of labeled training data, in particular taking into account the degree of variability in human activities. A possible solution is to leverage large scale online data repositories. This has been previously attempted with image and sound data, as both microphones and cameras are widely used sensing modalities. In this paper, we describe a first step towards the use of online, text-based activity descriptions to support general sensor-based activity recognition systems. The idea is to extract semantic information from online texts about the way complex activities are composed of simple ones that have to be performed (e.g. a manual for assembling a furniture piece) and use such a semantic description in conjunction with sensor based, statistical classifiers of basic actions to recognize the complex activities and compose them into semantic trees. Extraction of domain relevant information evaluated in 11 different text-based manuals from different domains reached an average recall of 77%, and precision of 88%. Actual structural error-rate in the construction of respective trees was around 1%.