How do you present AI and HCI projects to different audiences?

I agree with the points made in this article. I would like to suggest a couple more. When communicating technical information it helps to show that there's support for the idea from the scientific community. For example, the idea may have been published in a well known journal, could have a large number of citations and/or could have prestigious awards. This way even if the audience doesn't understand the idea completely, they are likely to accept it as it has the backing of the scientific community. Another point is that comparison with existing solutions should be shown. Having the real world or simulation results helps to show to what degree the idea outperforms or improves on the existing solutions.

If presenting to a non technical audience, I strongly recommend teaching by doing. Other than extreme basics about AIs and LLMs, a non technical audience doesn’t want to hear how it works - they can’t understand that. Stick to demoing a real product or prototype or proof of concept. Ideally, let people use it themselves to directly demonstrate value through experience.

In my work as a management consultant, I’ve learned that contextual relevance and empathy is paramount – whether I’m presenting to stakeholders in the tech sector or bridging the gap for those less familiar with AI and HCI concepts. You have to really understand your audience and why they should care about what you have to say. Content about AI must demonstrate how it will add value and help audience achieve their outcomes.

Make sure to use analogies with a non technical audience. For example: a LLM learns much like a self taught human. You give it all the information, it reads it all and processes it, then determines how to apply it. Unlike traditional machine or human learning, a teacher doesn’t specifically tell the computer HOW to achieve the desired result - the computer learns by inspecting millions of desired results, then figures out its own way to generate said results.

Demonstrating AI use cases is a powerful way to present to stakeholders. The cases need to show how AI has been used and the added value in tangible ways such as margin increase, cost reduction, risk mitigation, customer service improvement, access to new market, revenue generation ect..

Most current AI and HCI (human computer interaction) systems are not ethically trained. This means the people who produced the work that trained the robot are not getting paid and did not consent. I believe that we will have shifting opinions on this over time, as AI becomes more normalized and less of a culture shock to people. That said, I think it’s important to introduce ethics in the same paragraph as introducing the AI itself. I love Midjourney and ChatGPT. But they’re not ethical yet. It’s also important to address explainable AI / transparent AI. Unlike with ethics, this merely calls for transparency about training and about how the AI works.

Presenting AI and HCI projects effectively to different audiences involves tailoring message to their interests and backgrounds. For technical audiences, focus on detailed algorithms and methodologies, highlighting innovation and technical challenges. For non-technical stakeholders, emphasize real-world applications, benefits, and how the project addresses human needs, using relatable examples and plain language. During a "show and tell" session, consider using visually engaging demonstrations, interactive prototypes, and relatable scenarios to effectively showcase the AI and HCI projects, making it easier for the audience to understand the concepts and potential impact.

In addition, there are some behavioral science techniques that can help land presentations with different audiences. Eg. Anchoring involves using a reference point to influence people’s perception of value. In presentations, you can utilize anchoring by introducing a framework or visual and orienting the remainder of the presentation from that perspective. The Availability Heuristic refers to people’s tendency to rely on readily available information when making decisions. Personalizing AI use cases and examples based on the audience’s context is a way to leverage this. Being aware of such subtleties can strengthen the presentation and ensure it resonates with the audience.