iF Quarterly

AI & Us

How iFsters and our clients are incorporating generative AI in their work

It’s an exciting time to have access to the internet — (mostly) unfettered access to the wide, exponentially growing world of generative AI has rapidly changed the landscape of what is possible in industries ranging from healthcare to social impact to sustainability. Significant advances in large language models paired with a radical shift in accessibility has enabled individuals everywhere to unleash their creativity and explore generative AI’s potential with little to zero knowledge of the technology driving it.

Our work at iF spans a similar gamut of industry, and in most client engagements AI is bubbling to the forefront of our partners’ minds. Where does generative AI hold the greatest promise? Where does it introduce new risks? How can we accelerate our learning and experimentation?

Concurrent to client work and general knowledge sharing internally, the Digital Strategy & Innovation team recently concluded an organization-wide experiment to document how iFsters are using AI and for what purposes. The information below is just a small snippet of the many uses iFsters are finding for generative AI in their everyday work.

In the spirit of the theme of this edition of iF Quarterly, the headshots featured below were generated using a combination of iF headshots and Stable Diffusion, an open source generative AI tool. To address privacy concerns, Stable Diffusion was downloaded to an iFsters’ personal laptop, so no personal information — including imagery — was uploaded to the web.

How iFsters are using AI

Tynan Gable, Lead Strategist

“I use Bard to validate basic research findings, Otter.ai to synthesize meeting notes, ChatGPT to generate stories/written content as a starting point, and ChatPDF to summarize PDF content.”

Peter Arbaugh, Technical Product Manager

“I use ChatGPT constantly when I’m working with code. It’s a great tutor and helps me spot simple bugs, understand errors, and explore different approaches. It can also explain terms and help me figure out the right way to search for more information and access references online. I also find it to be a great starting point when starting a new project or learning about a new topic. It can give a high level overview which suggests further avenues for exploration.”

Olivia Rao, Project Manager

“I’ve used ChatPDF to help me sift through many long documents which has saved me time by summarizing research and answering my questions to help direct my time and attention to the parts that are most relevant to my work.”

Gentry Croley, Senior Strategist

“I always have ChatGPT open in my browser. I mostly rely on it to unearth and synthesize vast amounts of data and information, but I’ve also found the model to be a great ‘thought’ partner for more strategic efforts, such as crafting frameworks or determining how to present data given a set of parameters. I also use Otter.ai to transcribe and summarize notes during client meetings and user research interviews which saves me an enormous amount of time and effort.”

Malia Nakamura, Digital Marketer

“I use ChatGPT to help me generate marketing content fairly regularly, and I also use Read.ai to summarize meeting transcriptions and capture action items without being distracted by taking notes. I also use Photoshop’s new AI tools in my work as a photographer to remove/fill areas of images.”

Sam Bazant, Lead Strategist

“I mostly use ChatGPT for idea generation when I’m getting started with research for a project. It’s helpful to get a grasp of a new concept or market — the key trends, terms, and frameworks for thinking about products or major players. While it’s not always 100% accurate, it helps me get a jump start on how I might frame my research or where to look next.”

How we’re being responsible, safe & curious

Just like in our project work at iF, we take a learning disposition in our approach to exploring and testing new technologies — including generative AI (genAI). While there are plenty of reasons to be curious and excited about genAI, there are also reasons to be cautious. Some of the things that we’ve been paying close attention to are: Data accuracy, bias, and the human difference.

We’re doing a few things as we begin exploring genAI to maintain the highest standards of our work, delight clients, and shape the world for the better:

• Data Quality — When researching or analyzing data, we extract and validate sources, check relevancy and fidelity, and cross-check against multiple sources. If we can’t validate something, we do not include it in our work.
• Bias Potential — We engage in training and education on how to identify bias that genAI may include in responses, and strive to take an inclusive, equitable approach to our work, especially as relates to historically marginalized people and communities.
• Human Ownership — Because large language models (LLMs) are changing daily and ultimately are still machines, a skilled, trained, compassionate set of human eyes is needed, whether ideating, writing, or researching. We review, edit, and analyze genAI’s outputs for quality, relevancy, and accuracy.

We’re learning about this exciting new technology more and more, but we acknowledge our responsibility to do so in safe, ethical, accurate ways — for our staff, our clients and partners, and our communities.

How our clients are using AI

Likewise developed a ChatGPT plug-in, and they’re exploring further uses of generative AI to enhance their users’ experiences by coupling GPT with tens of millions of data points to make the process of discovering new content more engaging, personalized, and efficient.

Quill.org collaborated with CommonLit on a free new tool that allows educators to determine whether a text passage was created by humans or AI.