Sentiment analysis using AI

Continuing the discussion from Stakeholder consultation assisted by AI:

AI tools like ChatGPT can significantly enhance stakeholder sentiment analysis for policymaking by efficiently processing and analyzing large volumes of unstructured data from diverse sources.

Utilizing natural language processing (NLP) capabilities, these tools can assess opinions and emotions expressed in social media posts, news articles, blogs, public forums, correspondence and submissions.

This enables policymakers to gain a comprehensive understanding of stakeholder perspectives, identify key concerns, and monitor shifts in public sentiment over time. By incorporating these insights into the policymaking process, decision-makers can develop more responsive and targeted policies that address stakeholder needs, foster public trust, and promote broader social acceptance.

In addition, AI-driven sentiment analysis can help identify potential communication gaps, enabling policymakers to better engage with stakeholders and ensure that policy messaging resonates with the intended audience.

Sample AI prompts

What is sentiment analysis and how can it be used in policymaking?

What categories could be used to describe the sentiment of stakeholders towards a government policy?

Based on the following text, what is the sentiment of stakeholders towards [policy/issue]? Categorise the sentiment as “Supportive”, “Opposed” or “Neutral”. [Text]

What is the overall sentiment towards [policy/issue] among stakeholders in the following text, and are there any notable trends or patterns in their comments? [Text]

Analyze the sentiment regarding [proposal] and identify common themes and opinions within [text].

Based on the sentiment expressed in the following text, how might policymakers need to adjust their approach to address the concerns of the public? [Text]

What are stakeholders saying about [policy/issue] and how do they feel about it?

How this image was created using AI

The following prompt was used with ChatGPT (GPT-3.5):

Suggest some prompts to create an image to accompany the following website post. Talk about sentiment being good, bad or neutral. [post text]

One of the prompts suggested by ChatGPT was as follows (some minor edits were made):

Illustrate an cartoon style image that represents sentiment distribution across different stakeholders and topics. Show a heatmap with stakeholders and policy issues as rows and columns, with color intensity indicating the sentiment intensity. Use visual elements like a grid, color gradients, and sentiment indicators to symbolize sentiment distribution.

This prompt was then used to create an image using Image Creator from Microsoft Bing.