The 2024 Society for Applied Anthropology (SfAA) conference featured an insightful session on the intersection of anthropology and artificial intelligence (AI). Organized and moderated by Sonja Hodgson, the session brought together a diverse panel of experts, including Archana Shah, Lance L. Larkin, and myself. The discussion explored AI’s capabilities and limitations, ethical considerations, and the role of anthropologists in shaping the future of this rapidly evolving technology.

Limitations and Capabilities of AI

We acknowledged the broad scope of AI and its potential to transform various industries. However, we also emphasized the importance of understanding the limitations of current AI systems, particularly large language models (LLMs). As Archana Shah pointed out:

LLMs are next word prediction engines. As Matt pointed out, it may not necessarily understand what it is saying, but as long as we understand how it works and that’s all it is doing, we can use it very, very effectively to tell our stories much more strongly.

Highlighting the inherent challenges and opportunities within LLMs, Sonja Hodgson elaborated on the vital aspect of trust and user interaction with AI tools:

This is something we’ve worked a lot on, building our own tool and understanding. What are the trust limitations here and how can we, through good UX design, be better able to communicate those either limitations or communicate a level of trust.

We also discussed other limitations of LLMs, such as their lack of semantic understanding, potential for hallucinations, and lack of transparency in their decision-making processes, but we also emphasized that AI can have positive capabilities.

I highlighted the need for anthropologists to contribute to the development of AI, especially in areas like natural language processing and knowledge engineering, to ensure that AI is grounded in truth and represents diverse perspectives. I explained:

I’m particularly interested in how knowledge graphs can be joined with generative AI to ground LLMs in truth… We can work on that knowledge engineering to make sure that truth is modeled as accurately as possible for as many groups as possible to make sure those meanings really make sense.

Navigating these complexities underscores the imperative for an ethical framework in AI’s development and application, guiding us toward responsible innovation and inclusive progress.

Ethical Considerations

The session delved into the ethical implications of AI development and deployment. We emphasized the importance of human involvement in the AI development process and the need for ongoing monitoring and adjustments. Lance Larkin shared an example of a safety operator pulling the emergency brake to stop an AI vehicle from getting too close to pedestrians, illustrating the critical role of human oversight. He went on to say:

We need to unpack this black box so that we understand what that programming is doing in the background and also how it can go wrong, and how those recommendations that it makes are actually coming from a place that is not generating its own facts.

We also discussed the potential biases and misrepresentations in AI systems due to limited or skewed training data. I added to this, stating:

You have to consider things like bias and misrepresentation. You have to consider downstream effects. And so again, I think this is why we need a human in the loop… involved in the training and continuous improvement.

Adding to the conversation, Sonja Hodgson remarked:

“We need to be better advocates for those in the liminal spaces and understand how we can dig out outlier information from AI interpretations and predictions.”

This statement by Sonja Hodgson underscores the critical need for advocating on behalf of those in marginalized or overlooked segments of society, ensuring AI systems are developed and deployed in an inclusive, equitable manner. It complements the earlier points by illustrating the importance of a nuanced, comprehensive approach to AI ethics that goes beyond mere technical adjustments, advocating for a more holistic and human-centered perspective in AI development.

The Role of Anthropologists

The session also highlighted the unique opportunity for anthropologists to shape the development and deployment of AI by leveraging their expertise in understanding human behavior and culture. I encouraged anthropologists to move beyond critique and actively collaborate with other disciplines to build AI solutions:

We need to critique AI. We need to study it. We need an anthropology of AI. But more importantly, I think we need to prepare ourselves to not only use it in our work as anthropologists, whether in academia, government or business, but also, importantly, to contribute to building AI solutions.

We offered practical tips for anthropologists navigating the AI landscape, including:

  • Staying informed about the latest advancements in AI
  • Embracing design skills and data science to contribute to AI projects effectively
  • Advocating for ethical practices and diverse representation in AI development
  • Using AI as an accelerator and extension of oneself in research and storytelling

Lance Larkin emphasized the importance of anthropologists bringing their unique perspective and asking critical questions:

Even if you don’t want to get into AI, at least step back and still be an anthropologist and ask those anthropologist questions.

Archana Shah highlighted the potential for anthropologists to inform the development of AI by identifying areas of innovation and shaping the language and communication of AI systems:

There’s a lot of room for development. There are things that I think anthropologists in particular you’re trained to look for things that are unsaid, things that are not very obvious to you. And as you pointed out, asking the stupid questions to get to the kernel, and then expand out because in that kernel is where the biggest area of innovation lies.

As we consider these perspectives, it’s clear that the intersection between anthropology and AI offers fertile ground for innovation and ethical guidance.

Preparing for the Future

As anthropologists, we discussed the importance of adapting our methods and skill sets to engage with AI effectively. I suggested that anthropologists should embrace other disciplines to become even more valuable contributors within organizations:

Don’t shy away from other disciplines. I would encourage everybody to pick up some design and data science skills. Embrace the tools, because you will find that you can be a really good contributor within any organization, whether it’s academia, government, or for profit business.

I also discussed the need for more training and collaboration across academic disciplines to prepare anthropologists for the challenges and opportunities presented by AI. I emphasized the importance of providing students with practical skills and projects demonstrating their value to employers.

I would say to anybody in the audience who is in academia and particularly anybody that might be an administrator, we need more training for students to go out and work. More practical skills. They need projects that they come out with a portfolio they can show employers. And importantly, we need much more collaboration across academic disciplines. There’s no reason that anthropologists shouldn’t be taking data science classes. It would make us much broader and stronger as a discipline if we could do all of those things as individuals and collectively.


The SfAA session on anthropology and AI provided a thought-provoking discussion about anthropologists’ pivotal role in AI’s future, emphasizing the necessity for ethical engagement, cross-disciplinary collaboration, and the inclusion of diverse perspectives. As AI evolves, anthropologists must remain agile, enriching AI development with their unique insights. This commitment to ethical AI practices promises a more equitable future for society.

I extend my gratitude to my fellow panelists and the Society for Applied Anthropology for hosting this panel that underscored the significant contributions of anthropology to the realm of AI.