As public and user engagement practices are becoming more prominent with the AI-development space, AI-driven tools are also on the rise within the public engagement domain. To date, there are four broad categories of digital and AI-driven tools intended to support public engagement activities:

• Automated Synthesis
• Synthetic Participants & Facilitators (AI Chatbots)
• Synthetic Data
• Platform for Scalable Surveys

PAI’s Guidelines for Participatory Engagement primarily focuses on and prioritizes the application of community-empowering, qualitative approaches with an emphasis on building close working collaborations with people through direct (human) contact and frequent communication. This is a core principle used to develop a set of questions in order to analyze different public engagement tools that are coming to market:

• What parts of the public engagement process does the tool complement or replace?
• What aspects of human judgment and interaction are complemented or replaced?
• Can these AI-tools be developed and used in alignment with other Responsible AI principles (e.g., not violating user privacy or security, used in a transparent and consensual manner)
• Are human participants more or less empowered to make decisions about AI technology’s development and deployment through the use of these tools?

Automated Synthesis

Automated synthesis tools refers to AI-driven tools and platforms that can synthesize information on behalf of a human user. For example, a researcher may choose to enter a large corpus of written text (e.g., transcripts from 100 interviews) into a large language model (LLMs) to generate summaries of those interview notes. Generative AI tools may also be used to create materials for public engagement activities, such as creating a qualitative survey instrument (the series of open-ended questions the interviewer will ask Participants) based on parameters established by the user.

Synthetic Participants & Facilitators (AI Chatbots)

“Synthetic participants and facilitators” refers to AI chatbots that are designed to emulate a user persona (participant) or facilitate a public engagement activity (facilitator). Synthetic participants or users are AI-generated profiles that attempt to imitate a person or group programmed around a set of characteristics, in lieu of engaging with real people with those characteristics. These AI-generated personas can be used as the basis for chatbots so a researcher or engagement host can conduct in-depth qualitative interviews with different personas reflecting different public engagement groups. Quantitative surveys can also be “run” with a pool of AI-generated profiles to generate synthesized data. Synthetic facilitators are chatbots that can be programmed to ask participants a battery of questions, instead of having a human interviewer or session facilitator.

Synthetic Data

Synthetic data refers to artificially generated information that mimics the behavior and statistical properties of real-world data through algorithms, generative models, or simulations. It can be in the form of analyzing existing datasets to impute missing data (e.g., generating data points for groups of people who are underrepresented in the original sample) or through the use of synthetic participants which are programmed to mimic survey responses through specific user personas.

One common method for creating synthetic data involves using data seeds, which are initial examples or pieces of data used to guide the generation of additional synthetic data. Generating synthetic data is typically a recursive process, where data seeds are used to prompt language models to generate new synthetic data. Then, the synthetic instructions are given to more advanced language models to create responses that are used to train the target language model. Ultimately, the purpose of data seeds is to provide a starting point from which algorithms can generate new instances while maintaining specific characteristics or patterns.

Platforms for Scalable Surveys

While not always and explicitly AI-driven, there is a growing number of platforms designed to help tech developers connect with large audiences (1,000+ participants) to elicit their values and beliefs about AI, its development, deployment, and overall governance. While typically applied to governance questions (e.g., the broader vision of AI people believe will make it safe or broadly beneficial to society), other platforms exist to help connect developers with beta testers and other types of users to provide user experience feedback. These platforms are optimized for large populations, may focus on representative sampling approaches, can pose open-ended prompts (e.g., “I want to see…”), and may use real-time (algorithmically-processed) review of the responses to facilitate the discussion.