The Spark Hunter Survey was created to seek informed opinions about the nature, development, and impact of artificial general intelligence (AGI). Nearly all public media coverage of non-entertainment AGI comes from public statements of C-suite tech executives, tech scientists and engineers, or sci/tech academics and researchers. These three sectors all have their own substantial stakes—financial and professional—in public beliefs and opinions about emerging AGI: what it is and when it might appear. The people who are talking most frequently to all three sectors and cross-comparing and balancing opinions are sci/tech journalists. To the best of our knowledge, these journalists had never been asked systematically about their perspectives and opinions of this developing field.
The Spark Hunter Survey was conceived by AGI Ethics News, developed by scientific and technical journalists/writers and philosophers, reviewed by an experimental psychologist and a psychometrician, edited by Perplexity, and executed by Fighter Steel Productions LLC, the publisher of AGI Ethics News. Specifically, we assembled a list of approximately 400 of the top English-speaking sci/tech journalists in the world and recruited participants from that list. Thirty-one journalists completed the survey, and were each paid $50 for completing the survey, except for one journalist who declined payment.
The survey data was then forwarded to Professor Sheldon Solomon at Skidmore College, working with Findlay Tyler and Brian Newton, students in The Human Motivation Lab, to analyze and interpret the survey findings in order to produce this preliminary report. The lab group has previously contributed to research investigating the prospective benefits of AI-assisted psychotherapy and has more recently been reflecting on a recent MIT study (Kosmyna et al., 2025) demonstrating that heavy reliance on ChatGPT reduced neural connectivity and impaired memory for one’s own writing, as well as Aradhay Sharma’s (2025) Gen Z x AI: Thrive Together account of how artificial intelligence is reshaping education, work, and daily life, with recommendations for using AI ethically and effectively to build skills, innovate, and prepare for future careers.
The survey included 39 Likert-type items, 14 open-ended/free response questions, 3 demographic questions, and 4 administrative questions such as consent, follow-up interest, and payment/contact details.
On the demographic front, respondents were from the United States (58.1%), the United Kingdom (16.1%), India (9.7%), Mexico (3.2%), Macedonia (3.2%), Kenya (3.2%), South Africa (3.2%), and Australia (3.2%). The number of years respondents had worked as journalists ranged from 1 to 50 years (mean = 13.1 years; SD = 9.5). The majority of respondents (58.1%) identified as freelancers contributing to multiple outlets, with the next largest group affiliated with online-only news platforms (19.4%), followed by magazine and newsletter journalists (12.9%), national newspaper/wire service journalists (6.5%), and local newspaper journalists (3.2%).
We then divided the Likert-type items and open-ended questions into four factors in order to streamline the findings conceptually, rather than reporting the specific responses to each individual question. For each factor, we present general findings along with specific results for representative Likert-type questions and open-ended responses.
Factor 1: Perceived pace and significance of AGI
This factor explores whether respondents view AGI as a consequential and potentially rapidly advancing development rather than a distant or purely speculative concept. Items in this cluster concern timelines, perceived precursors to AGI, likely impacts, and the degree to which AGI deserves journalistic attention. Together, these questions measure a form of technological urgency: not necessarily certainty that AGI is imminent, but a belief that it is important enough to warrant active attention.
The representative question for this factor is: “AGI development is progressing faster than most people realize.” Here the most common answer was Agree, selected by 41.9% of respondents; 35.5% chose Neither agree nor disagree, while 9.7% selected Disagree, 9.7% selected Strongly disagree, and 3.2% selected Strongly agree. This pattern suggests a moderate sense of urgency, with the sample leaning toward the view that AGI is advancing faster than public understanding. The dominant orientation is that AGI should be carefully monitored, even if respondents are not ready to endorse the strongest claims about immediacy.
Typical open-ended responses from an item asking respondents to define AGI for a layperson:
“AGI is a hypothetical type of AI that can match or surpass human performance across any cognitive tasks. Unlike narrow AI, it excels across many intellectual domains.” This response is representative because it frames AGI as a broad, cross-domain capability explicitly contrasted with narrow AI, reflecting the seriousness and conceptual weight respondents often brought to the topic.
“AGI is a machine that can reason across tasks rather than being limited to one area.” This short definition reinforces the same broad distinction between general and narrow intelligence.
“It would be an AI that can learn and apply knowledge across multiple domains like a human.” This emphasizes flexibility and generalization rather than narrow task performance.
Factor 2: Governance, ethics, and responsibility
The second factor reflects respondents’ views about whether AGI requires early governance, ethical safeguards, and institutional responsibility. Questions in this area deal with alignment, safety, oversight, responsibility, and the importance of proactive intervention.
The representative question for this factor is: “How important are proactive measures to ensure the responsible and ethical development of AGI, such as investing in safety/alignment research and establishing strong ethical guidelines/standards for developers?” On this item, 54.8% of respondents selected Critically important, and 38.7% selected Very important. Only 3.2% selected Moderately important, and 3.2% selected Slightly important. This distribution shows the strongest consensus among the four representative questions. With 93.5% placing the item in the top two categories, the sample demonstrates overwhelming support for anticipatory governance and ethical oversight. Even respondents who may disagree about AGI’s feasibility or timeline seem to converge on the judgment that passive waiting would be irresponsible.
Typical open-ended responses asking respondents for one piece of advice to policymakers:
“Ensure that AGI development is governed transparently, with public input and independent oversight.” This statement captures the broader moral orientation of the governance factor by emphasizing transparency, public accountability, and independent oversight rather than leaving AGI development solely in the hands of private actors.
“Do not regulate imagined superintelligence while ignoring present power asymmetries, labor impacts, and data governance.” This adds a more concrete policy critique and extends the governance concern to current harms.
“Stop waiting for AGI to arrive before governing it.” This response captures the survey’s recurring emphasis on anticipation rather than reaction.
Factor 3: Conceptual clarity and definitional confidence
Questions in this cluster concern whether respondents believe AGI is sufficiently conceptually coherent to support meaningful discussion and evaluation. It includes questions about whether AGI is well-defined, how it differs from narrow AI, whether current systems qualify, and whether core concepts such as alignment are understood.
The representative question for this factor is: “Do you believe the term ‘Artificial General Intelligence’ itself is well-defined and understood within the scientific community?” On this item, 46.9% of respondents selected No, 28.1% selected Yes, and 25.0% selected Unsure.
Typical open-ended responses to “What do you foresee as the biggest challenge in verifying that AGI has been achieved?”:
“Distinguishing true understanding from sophisticated mimicry.” This response expresses a clear boundary between narrow AI and AGI, indicating how respondents often grounded their judgments in distinctions they regarded as conceptually meaningful rather than merely rhetorical.
“Defining general intelligence precisely enough.” This response focuses directly on the conceptual problem of criteria rather than only on performance.
“Lack of transparency in AGI systems.” This captures the verification problem as one of epistemic access rather than only definitional disagreement.
Factor 4: Media framing and the journalistic role
The fourth factor captures beliefs about how journalism should cover AGI and whether current media discourse is adequate. This includes questions about balance, objectivity, critical reporting, differentiation between narrow AI and AGI, and the role journalists ought to play in public debate. Because the sample is composed of journalists, this factor reveals an especially important layer of professional self-understanding.
The representative question for this factor is: “What is the primary role journalists should play regarding AGI discourse?” The most common response was “Primarily scrutinize claims made by developers and proponents,” selected by 41.9% of respondents, followed by “Primarily explore potential societal impacts (both positive and negative)” by 19.4%. Then 9.7% respondents selected “Primarily inform the public about developments,” 9.7% selected “Primarily advocate for responsible development and ethical guidelines,” 9.7% chose “Maintain strict neutrality, reporting all sides equally,” and 6.5% selected “Other.”
Typical open-ended responses from asking what is most missing from public and media discussions of AGI:
“Realism. Public and media coverage is dominated by oversimplified or sensationalised narratives, rather than reflecting the true progress, challenges, risks, and benefits.” This representative response criticizes sensationalism and oversimplification while calling for a more realistic account of both risks and benefits. It aligns closely with the closed-ended result showing that the preferred journalistic role is critical scrutiny.
“How do we ensure it’s safe?” This concise question reflects the field’s strong concern with risk and verification.
“What are its true capabilities?” This also captures the journalistic impulse to probe claims rather than reproduce them.
Summary and conclusions
In sum, respondents generally view AGI as meriting close journalistic attention, overwhelmingly support proactive governance and ethical safeguards, and favor a journalistic role rooted in skepticism and public accountability. They define AGI in broad cognitive terms, ask for transparency and oversight, criticize sensationalist coverage, and orient journalism toward testing claims rather than amplifying and transmitting them.
Factor 1
When, if ever, do you believe AGI will be developed?
The potential benefits of AGI are often understated in public discourse.
Factor 2
Concern: AGI developed by a small number of powerful entities (corporations or nations)
Concern: Lack of transparency by AGI developers
Factor 3
AGI will be capable of generating creative works indistinguishable from humans within 10 years
AGI will be integrated into critical infrastructure (e.g., power grids) on a large scale within 10 years
Factor 4
AGI is more likely to help alleviate global environmental crises than exacerbate them
Concern: AGI’s impact on global inequalities
AGI contributing to major scientific breakthroughs (e.g., in medicine, materials science, or climate modeling) that were previously intractable.
