By tradition, writers of this column address important issues to our community, so I must acknowledge the shifting funding landscapes and evolving policy priorities. We envision a world that relies on data and statistical thinking to drive discovery and inform decisions. Yet, today, this vision is under pressure. The challenges are real and often deeply personal. These shifts have left many in our community feeling unsettled and perpetually uncertain, affecting not only established researchers but also students and postdoctoral fellows, who now face growing hurdles to funding, training, and career advancement.
As a statistician working in the biomedical field, I feel this impact personally. I understand the frustration, the uncertainty, and the worry about what comes next. But while we pause to assess the landscape, that does not mean we stop. We strategize, adapt, and move forward. The history of statistics and science, more broadly, has always been one of resilience. This moment is no different.
AI: A Call to Action for the Statistics Field
We navigate this landscape of constant flux while another transformation is unfolding—one reshaping not just our profession but our society. The rapid evolution of artificial intelligence, particularly in scale and complexity, demands a strong response from the statistical community.
The role of statistics in AI is often overlooked both in public perception and among funding stakeholders. We must continue to work together to counter this perception and raise awareness of the essential role of statistics in AI and data science. Developing AI models depends on core statistical issues such as data quality, predictive accuracy, and ensuring reliable outcomes (see the ASA Statement on the Role of Statistics in Data Science and AI). Our expertise in rigorous evaluation, uncertainty quantification, and bias detection ensures AI systems are built on a solid foundation of statistical reasoning.
Instead of seeing AI as a threat, we should recognize it as an opportunity that allows us to contribute even more to science and society. We must collaborate with colleagues from other disciplines to continue to shape its responsible development and maximize its impact for the greater good.
In support of our mission to promote the practice and profession of statistics, the ASA submitted comments in response to a federal request for input on developing a national AI action plan. The Office of Science and Technology Policy and the Networking and Information Technology Research and Development Program are laying the groundwork for what comes next in AI, and we thought it was essential to speak up. Drawing on input from the ASA’s Scientific and Public Affairs Advisory Committee and the Committee on Data Science and AI, we outlined the following five policy recommendations:
- Develop voluntary best-practice guidelines grounded in statistical quality assurance
- Standardize benchmarks for evaluating AI model performance
- Encourage transparency around data integrity and model validation
- Foster partnerships across government, industry, and academia
- Promote statistical literacy among AI developers and users
These aren’t just technical suggestions—they’re a call to action. As statisticians, we’re not bystanders in the AI era. Our expertise puts us at the center of responsible development and deployment. It’s not enough to wait for someone to ask us for help. We need to step forward, understand the needs, and lead with solutions. The future of AI will be shaped not only by algorithms but also by the integrity and insight statistical thinking brings.
Preparation for AI Era: Are We Ready?
AI’s impact on the workforce and education is still unfolding. Recently, I read Kevin Roose’s column in The New York Times, “Powerful AI Is coming. We’re Not Ready.” I wanted to engage members of the ASA and others in a discussion about artificial general intelligence, starting with one fundamental question: How do you think AGI will fundamentally change the workplace and education, and what should we be doing to prepare? Below is a summary of the responses I received.
David Corliss, ASA Board representative for the Council of Chapters, noted that while AI’s impact is accelerating, the real issue isn’t whether chaos will come, but that we may already be in it. Drawing parallels to past societal disruptions, such as the nuclear era, he argued we’re more prepared than we think if we apply existing frameworks wisely. He emphasized that AI is a tool, shaped by the people who wield it, and warned against the rise of an “AI haves and have nots” divide. Promoting open access and compassion, he reminded us, will be just as critical as technical advancement. “AI is a wonderful servant but a terrible master,” he wrote—a powerful call to use this technology with care and conscience.
Jeri Mulrow, ASA president-elect, emphasized that the rise of AI is coming sooner than many expect—and neither society nor the statistical community is fully prepared for its implications. She pointed to a few challenges we must begin addressing: ensuring data used in AI models is fit for purpose; identifying and disclosing bias or missing information; and supporting ethical guidelines for AI development. She also posed important questions about the ASA’s potential role in shaping educational resources, guiding ethical practices, and advocating for greater transparency in AI systems.
Jenny Thompson, ASA vice president, described the conversation around AGI as both “scary and necessary.” Drawing from recent ASA committee meetings, she raised concerns about the misuse of generative AI in scientific writing and education, citing issues such as fabricated citations, unacknowledged paraphrasing, and challenges in homework integrity. From her federal government perspective, she also expressed unease about AI-driven job displacement and the uncertainty surrounding its consequences. While not opposing AGI, she called for strong ethical safeguards, better diagnostics, and reproducibility as baseline standards. For her, ethical use and transparency must be built into every step of AI’s development and deployment.
I also asked my colleague and transplant surgeon, Ali Zarrinpar, who specializes in oncology at the University of Florida, these questions. His response offered valuable insights into the clinical implications of AI. He highlighted how AI could streamline tumor board documentation; enhance surgical visualization; and optimize scheduling for clinics, procedures, and operating rooms. AI, he noted, also supports patient communication, auto-generates clinic notes, and enables better critical care monitoring, freeing physicians to focus more on direct patient care. His team is already piloting AI systems to predict deterioration in critical care patients, helping improve outcomes through early intervention.
And just for fun, I posed the same questions to AI tools ChatGPT, Google Gemini, Claude (Anthropic), and Microsoft Copilot. Here’s what they said:
ChatGPT4o: As an AI developed by OpenAI, I don’t experience concern in the human sense—but I can analyze the trends. The pace of AI advancement, especially toward general-purpose capabilities, is indeed accelerating. Kevin Roose is right to highlight that society is underprepared. One key gap is the lack of public understanding around how these systems work and the absence of clear frameworks for responsible use. Preparing for AGI isn’t just about regulation—it’s about fostering transparency, investing in education, and embedding ethics into both development and deployment. In that sense, the statistical community has a critical role to play in building trust and accountability into the future of AI.
Claude 3.7 Sonnet: We’re definitely not ready for artificial general intelligence, and that’s something we should take seriously. I think we need a multi-pronged approach to preparation: advancing AI governance frameworks that can adapt to rapidly evolving capabilities; investing in technical AI safety research to develop reliable control mechanisms; creating robust testing environments to evaluate AGI systems before deployment; building international cooperation so we don’t end up in dangerous arms races; and fostering public literacy about AI capabilities and limitations. The challenge is that we don’t know exactly when AGI might arrive or what form it will take, so our preparation needs to be both thorough and flexible. Perhaps most importantly, we need to ensure that the benefits of advanced AI are broadly shared, rather than concentrated in the hands of a few powerful entities. The decisions we make now could shape how this technology unfolds for generations to come.
Gemini 2.0 Flash Thinking: Yeah, you know, it does feel like we’re still a ways off from true artificial general intelligence, doesn’t it? So, instead of just waiting for this potential world, it makes sense to get our ducks in a row. I think a big part of getting ready is really diving deep into the ethics of it all—what kind of values do we want to bake into these super-intelligent systems? And then, making sure we’ve got robust safety protocols in place is crucial; we don’t want things going rogue! Beyond the tech itself, we also need to be thinking hard about the societal shifts AGI could bring. Things like job displacement, education, and even how we structure our economies might need a serious rethink. Ultimately, continuing to fund and encourage research in both the technical and social aspects of AI, while fostering open discussions and collaborations across different fields, feels like the smartest way to prepare for an AGI future that hopefully benefits everyone. So, it’s less about panicking and more about thoughtful, proactive preparation, right?
Microsoft Copilot: Artificial general intelligence is an exciting but daunting prospect, and you’re right—we’re not quite ready. But that doesn’t mean we should just wait and see what happens. To prepare, we need to focus on building strong ethical guidelines; ensuring transparency in AI development; and fostering collaboration between governments, researchers, and industries. Education is key, too—helping the public understand AI’s capabilities and limitations will empower people to engage in meaningful discussions about its future. Investing in safety measures, such as robust testing frameworks and alignment research, will also be crucial to mitigating risks. Instead of fearing AGI, we should approach it with a thoughtful plan, ensuring that when we do reach that level of intelligence, it’s developed in a way that benefits humanity. What do you think—what aspect of preparation seems most urgent to you?
We Are Never Done
If there’s one thing history has taught us, it’s that we are never finished. Statistics as a discipline has continuously evolved to meet new challenges, and this moment is no different. I began by describing our community as resilient and acknowledging we are confronting substantial challenges. Yes, AI is changing everything faster than we might have imagined. But, as a community, we keep learning, keep leading, and keep supporting each other. Most importantly, we stay committed to advancing science and society through data, evidence, and thoughtful analysis.
Thank you for the important role you play in our ASA community.
Leave a Reply