The exact text we saved

This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.

TRUVACE RECORD VERSION
record: TRV-2026-2701
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-11-12T17:03:13.048404Z
status: published
lens: p_space
sector: climate
headline: Public perceptions of AI energy consumption: Awareness, responsibility attribution, and strategies
dek: Abstract Generative artificial intelligence (GenAI) is widely used in the gig economy and electronic markets. However, its progress incurs substantial environmental costs due to its high and constantly increasing energy consumption. As inference—the process by which GenAI generates output—accounts for a large part of this demand, individual usage patterns are becoming a key factor in understanding the overall energy footprint and a crucial prerequisite for developing effective educational strategies and policies…
gain_title: (none)
problem_title: In a survey of 1,080 US adults, respondents gave users 22.18 of 100 responsibility points for generative AI's energy use versus 38.82 for household appliances, and only 5.7% of suggested fixes involved optimizing how AI is used.
trace_subject: (none)
gain_reading: (none)
gain_evidence: (none)
problem_reading: In a survey of 1,080 US adults, respondents gave users 22.18 of 100 responsibility points for generative AI's energy use versus 38.82 for household appliances, and only 5.7% of suggested fixes involved optimizing how AI is used.
problem_evidence: For AI, responsibility is more diffused and users are seen as less responsible, while developers and policy-makers are seen as more responsible.
quick_read: A survey of 1,080 US adults, built to match the US Census on sex, age and ethnicity, found that people largely understand which AI tasks use more energy but put little of the responsibility on themselves. Asked to split 100 points among five actors, respondents gave users 22.18 points for generative AI's energy use and 38.82 for household appliances.

The shift went to developers (28.69 points for AI versus 17.72 for appliances) and policy-makers (18.92 versus 13.34). Vendors were treated about the same in both cases (25.76 versus 27.15, not a significant difference). The authors read this as responsibility spreading across many actors when computing happens in remote data centers.

When asked what they could personally do to save energy with generative AI, 44.9% of coded answers were about using it less, abstaining or switching to non-AI tools. Only 5.7% named choosing a provider, model or output type that uses less energy. The authors note that inference accounts for roughly 60% of AI's energy demand, so how people use these tools matters.
limitation: Responses come from a Prolific crowdsourced panel, and the strategies people proposed were not tested for their real effect on energy use.
tag: Evidence-backed problem
key_points: Participants ranked five AI tasks by energy use with about 3.46 swaps needed to reach the correct order, versus 3.06 swaps for five household tasks (p = .037), a small gap in accuracy. | Respondents correctly sensed that tasks such as image generation use more energy than question answering, which the authors take as basic awareness of AI energy demand. | Policy-makers received 18.92 responsibility points for AI versus 13.34 for household devices, and developers 28.69 versus 17.72. | Users received 22.18 points for AI versus 38.82 for household devices (Cohen's d = −0.65); 'no one' got 4.37 points for AI versus 2.69 for devices. | Of the strategies respondents proposed, 33.5% were reducing AI use, 8.2% abstaining entirely and 3.2% using non-AI alternatives, together 44.9% of responses. | Provider choice (1.3%), model choice (2.1%) and modality choice (2.3%) together made up 5.7% of responses, while 9.9% of answers were 'not sure'. | Higher self-reported AI proficiency predicted suggesting model choice (B = 0.43) and prompt optimization (B = 0.26); more frequent users were more likely to propose cutting use.
rundown: Data came from an online survey on the Prolific platform. Of 1,098 respondents, 18 were dropped for saying they had not answered seriously, leaving 1,080 aged 18 to 89 (mean age 45.94). Respondents were paid £1.24 for a median of nine minutes. Authors Michael Klesel and Uwe Messer framed the work with construal-level theory, which holds that things felt as psychologically distant are thought of more abstractly.

Awareness was tested with two drag-and-drop ranking tasks, each covering eight hours of use. The AI tasks were image generation, image captioning, speech recognition, question answering and text classification, ordered using Luccioni et al.'s per-task energy estimates (477 Wh for image generation versus 0.5 Wh for text classification). The household tasks ran from a space heater (10.6 kWh) down to a cable modem (0.048 kWh), using US Department of Energy figures. Correlations with the correct order were 0.37 for AI and 0.46 for household tasks.

Responsibility was measured with a modified dollar-metric scale: 100 points split among policy-makers, vendors, users, engineers and no one, asked separately for AI and for appliances. Differences for policy-makers, users, developers and 'no one' stayed significant after Bonferroni correction. The authors interpret the move away from users as consistent with a cloud-based service feeling remote.

Open-text answers on personal energy-saving strategies were coded into seven categories using a theory-derived framework and an LLM ensemble. Logistic regressions showed that men were more likely to suggest modality choice (B = 0.79), and that younger and more proficient respondents suggested model choice. People who used AI less often were more likely to say abstain. The authors call prompt optimization and model selection valuable but note most respondents did not think of them.
sources:
- peer_reviewed | Electronic Markets | https://doi.org/10.1007/s12525-026-00908-w | 2026-06-25
prev: 0000000000000000000000000000000000000000000000000000000000000000

SHA-256 hash

935550d5ed98fbf50a1b133d2985d72b8503d1548c382bd2b42a3b5e555f789a

Previous version’s hash

0000000000000000000000000000000000000000000000000000000000000000