Artificial intelligence (AI) is a broad umbrella term that refers to a range of technologies that have been deployed for scientific, medical, commercial, and surveillance purposes. In my opinion, critiquing AI as if it is one thing is not particularly fruitful. Like computers, AI can be deployed for ethical and nonethical purposes, and it can be designed, executed, and utilized in more and less ethical ways.
There are three major concerns I have with AI.
The first concern I have is the major environmental concerns. In the United States, in particular, data centers are being constructed at a rapid pace without the necessary environmental regulation and technology deployment to limit energy use, water consumption, and pollution risk for neighboring communities. AI is also increasingly being utilized for commercial purposes that provide little societal benefit but use a massive amount of natural resources. This includes the most resource-intensive applications of AI such as image and video generation. Many of the most serious environmental issues with AI could be addressed by governments regulating data center construction and limiting/penalizing AI’s uses for less socially beneficial purposes.
The second concern I have are social concerns. AI algorithms, particularly for large-scale language models like ChatGPT and Claude, were trained with the nonconsensual use of a large amount of intellectual and artistic property. I believe there is a sanctity to the human creative process, and it is important that we do not allow these machines to steal the vital role of art in culture and the livelihoods of artists and others working in the creative economy. Many of these models were also trained with unfair labor practices, subjecting workers in the Global South to intensive and often horrific moderation jobs that could have been avoided with more ethical business practices. There are also serious issues of surveillance and data privacy/ownership that exist with certain applications of AI.
The third concern I have is with the employment effects of job automation. AI is projected to result in large amounts of unemployment in the next 5 to 10 years, affecting workers in a range of professions. If the government does not act responsibly to provide unemployment insurance, job retraining, and universal basic income, this is likely to result in a public health crisis.
When it comes to individual users of AI, I believe people lecturing other people about using AI is similar to lecturing your neighbor for keeping their lamp on when there is a coal plant operating next door. Sure, we all do have some responsibility for our individual choices, but individual users of AI have a relatively small impact compared to the corporations deploying it, the corporations developing it and profiting off of it, and the government policymakers responsible for regulating it.
On a personal level, I intentionally limit my use of AI tools in my creative and professional work, with a few exceptions. This use of AI aligns with my values of reducing environmental impact, avoiding displacement of creatives, and not financially supporting AI companies, while leveraging lower-impact uses of AI for intentional purposes. This includes occasional use for mental health processing, as a tool to help me work through brain fog, synthesize disoriented thoughts, and hear my thoughts reiterated back to me, as a tool for audio transcription, as a tool for business/innovation feedback, as a tool to receive encouraging feedback when working alone, for limited editing assistance on routine documents, and code review. I specifically do not use AI for image or video generation, and I do not personally pay for AI services (e.g., Claude, ChatGPT). I have made a commitment to not use generative AI in my publication writing, fiction writing, and other art, a commitment I am reiterating publicly with this note.
The landscape of data analysis and quantitative research is currently changing because AI tools are becoming increasingly integrated into the methodological norms of public health, social science, and other disciplines. In these fields, AI is not replacing research design, statistical knowledge, theory, or subject-matter knowledge. It is however becoming an important technology for quantitative data analysis, code writing, and software design. Methodological approaches are shared across disciplines, and already, AI-based methods and AI-assisted workflows are becoming a standard in many corners of the computational sciences. While I continue to code manually for the projects that I lead, I use AI tools when they are part of the workflow of teams I am collaborating with and it is necessary for collaboration.
I do believe there are many important and likely world-changing applications of AI in science, medicine, public health, and other fields, and in my view, choosing not to engage with AI at all is similar to choosing not to use a computer. There are in fact significant ethical problems with computers, but it is challenging to contribute to or engage with society today without using a computer. Increasingly, most apps and systems are integrating AI, so whether you are aware of it or not, AI has become a significant part of our day-to-day lives.
I don’t think it will work to just try to make AI go away. I think we need to recognize that it is here to stay, that it needs to be very heavily regulated by policymakers, and that it has the potential of delivering societal benefits if we insist on this through policy.
I will continue to update this position statement as disciplinary norms evolve and as my own policies develop and change.
