Using AI avatars for personalised disaster relief.

In partnership with

Dear reader,

By now we know that artificial intelligence (AI) has the potential to revolutionise disaster relief. 

In this edition of the Wag The Dog newsletter, I'll be looking at one particularly promising application of AI in this context; the use of AI-generated avatars to provide personalised support to people affected by a disaster.

While the concept of AI avatars may seem too far-fetched or even controversial to some, I believe that when professionally managed and trained, they can offer valuable support during emergencies.

These avatars can be designed to respond to the individual needs and circumstances of each person and provide customised help and support in times of crisis.

This human-centred approach to disaster relief would take into account the diversity of experiences and challenges faced by those affected and would make a more empathetic and effective response possible.

So, let’s dive in, and I look forward to reading your feedback!

Table of Contents

• The need for personalisation in crisis response

• How do AI avatars work?

• Practical use of AI avatars in emergencies

• Local empowerment through AI avatars

• The risks and limitations of AI

• A hybrid approach: AI and human expertise

• Moving Forward: Balancing technology and empathy

• References and further reading.

• Let’s meet!

• Sponsor

• What I am reading/testing/checking out:

The need for personalisation in crisis response

In the past, blanket disaster relief models have overlooked the different needs of affected populations.

One example is Hurricane Katrina (2005), where marginalised populations, particularly low-income and minority groups, had difficulty accessing aid due to systemic barriers. (See my previous post about this as well).

AI avatars could have bridged these gaps by tailoring communications and resources to the specific needs of the community. For example, they could have provided information in multiple languages and adapted their delivery to fit cultural realities.

How do AI avatars work?

AI avatars, which are based on natural language processing and machine learning, analyse real-time data such as social media, local reports and satellite imagery to respond to people’s needs. They are tools for efficiency and also for empathy.

Imagine an elderly person trapped in a scenario after an earthquake. An AI avatar could give age-appropriate, understandable instructions, guide the person to safety and provide emotional support during the ordeal.

Research into AI-driven support tools¹ shows that these are able to analyse real-time data and give support in an emotionally appropriate way faster than traditional methods.

Practical use of AI avatars in emergencies

Different strategies can be used to ensure that AI avatars reach people effectively in crisis situations:

• 📱 Mobile devices: AI avatars can be accessed via smartphones and tablets to utilise widespread mobile technology.

• 💻 Public information kiosks: The use of avatars in community centres or emergency shelters offers real-time help to those who do not have their own devices.

• 🌐 Web platforms: Embedding avatars in websites increases outreach, even in remote areas.

• ⚠️ Emergency alert systems: Integrating avatars into alert systems enables personalised, interactive responses.

• 🗣 Voice-controlled assistants: Avatars on platforms such as Alexa or Google Assistant⁵ enable hands-free interaction.

• 👨‍👩‍👧‍👦 Community partnerships: Collaboration with local organisations ensures vulnerable populations have access to this support.

Using these methods, AI avatars can provide vital, personalised assistance in emergencies, ensuring help gets to where it is needed most.

Local empowerment through AI avatars

Local leaders play a central role in effective disaster relief, and AI avatars can amplify their efforts. In the 2018 tsunami disaster in Indonesia, AI could have helped local leaders coordinate the distribution of relief supplies through existing networks and minimise red tape.

In regions where mistrust of external aid is high, culturally sensitive AI avatars can serve as trustworthy intermediaries. These tools can be programmed to recognise local customs and languages, enabling a more inclusive response.

This aligns with the knowledge that incorporating local insights leads to better outcomes in emergency response.

The risks and limitations of AI

Despite their benefits, AI systems carry inherent risks, especially around biases in data collection. If an AI avatar is built on flawed data, it could perpetuate inequalities, offering insufficient support to underrepresented communities.

Ethical development and rigorous human oversight are critical to ensuring these systems do not exacerbate the very challenges they seek to solve.

Research on AI-driven emergency frameworks³ emphasises the need for data accuracy and human input to minimise bias and ensure equitable responses.

A hybrid approach: AI and human expertise

AI avatars should not replace human responders, but complement them. In the Australian wildfires of 2020, AI systems analysed satellite imagery to predict fire patterns and allocate firefighting resources² .

However, it was human intuition that adapted to the rapidly changing weather conditions on the ground.

A similar balance can be achieved with AI avatars in disaster relief. They can process large amounts of data to suggest responses, but the final decisions need to be made with human expertise, especially in culturally or emotionally sensitive situations.

Moving Forward: Balancing technology and empathy

The future of disaster relief lies in the combination of technological innovation and empathy. AI avatars have the potential to transform disaster relief by personalising support, promoting trust and empowering the population.

However, we must be mindful of ethical concerns and ensure that these systems complement, rather than replace, human assistance.

The experience of AI-driven systems such as the Mpox screening tool for low-resource settings⁴ shows that even the most advanced technology needs to be developed with empathy, inclusion and accessibility in mind.

Disaster relief is not only about delivering aid, but also about building trust, resilience and long-term capacity in affected communities.

By placing these values at the centre, we can ensure that AI avatars compassionate, providing a more equitable, responsive and humane approach to crisis management.

What do you think? Would you trust an AI Avatar to assist you through an emergency?

🎧 Do you listen to podcasts? This newsletter is now available in audio format on Google Podcasts, Spotify, Stitcher, Deezer, Listennotes and many more.

References and further reading.

1  Choi Gyun. (2021). A Study on Wellbeing Support System for the Elderly using AI. Journal of Convergence Information Technology; https://www.semanticscholar.org/paper/A-Study-on-Wellbeing-Support-System-for-the-Elderly-Gyun/96db150f6fef2c7efb38427e7a67b55a71ef9e67#paper-topics

2  Bakhtiyor Meraliyev, & Kurmangazy Kongratbayev. (2020). Applying machine learning models for predicting forest fires in Australia and the influence of weather on the spread of fires based on satellite and weather forecast data. Proceedings of International Young Scholars Workshop; https://www.semanticscholar.org/paper/Applying-machine-learning-models-for-predicting-in-Meraliyev-Kongratbayev/b80d6bf77a067d5be24d4e71ed052368b7fb7a00

3  Li, Z. (2024). Leveraging AI automated emergency response with natural language processing: Enhancing real-time decision making and communication. Applied and Computational Engineering; https://www.semanticscholar.org/paper/Leveraging-AI-automated-emergency-response-with-and-Li/b1a97860dd5ebbe4c80cd412c0bf08c1ef0fe6f4

4  Kularathne, Y., Janitha, P., & Ambepitiya, S. (2024). Mpox Screen Lite: AI-Driven On-Device Offline Mpox Screening for Low-Resource African Mpox Emergency Response. ArXiv.org. https://arxiv.org/abs/2409.03806

5  Do, V., Huyen, A., Joubert, F. J., Gabriel, M., Yun, K., Lu, T., & Chow, E. (2022). A virtual assistant for first responders using natural language understanding and optical character recognition. Defense + Commercial Sensing; https://www.semanticscholar.org/paper/A-virtual-assistant-for-first-responders-using-and-Do-Huyen/f596981a93f3ddd91f7e1d11ad9d67ef0ec7d851

Let’s meet!

Here are the events and conferences I'll be speaking at. If you're around, feel free to message me and we can meet up for a coffee or a Negroni.

🌍 International Public Relations Association (IPRA) Thought Leader Webinar: Using AI for Crisis Communications, Thursday 12 September, online.

🇺🇸 Al in PR Conference + Bootcamp, 17-18 October 2024, Chicago, USA

🇬🇧 Crisis Communications Boot Camp, 4-5 November, London, United Kingdom

🇺🇸 International Association of Emergency Managers (IAEM) Annual Conference, 7 November, Colorado Springs, USA (remote/virtual).

🇳🇿 Emergency Media and Public Affairs (EMPA) conference, 7 November, Wellington, New Zealand (remote/virtual)

🇧🇪 AI in PR Boot Camp II, 20-21 February 2025, Brussels, Belgium

Sponsor

The Daily Newsletter for Intellectually Curious Readers

• We scour 100+ sources daily

• Read by CEOs, scientists, business owners and more

• 3.5 million subscribers

Sign up today!

What I am reading/testing/checking out:

• News: OpenAI plans to release 'Strawberry' for ChatGPT in two weeks

• Webinar: September 16. How hate spreads - The real-world consequences of disinformation.

• Article: Amazon strengthens its emergency relief efforts

• Paper submission. You can now submit your paper for the International Crisis and Risk Communication Association (ICRCA) conference.

• Podcast: Not If, But When - Learning the lessons of the last

  pandemic

• Webinar: October 17. Artificial Intelligence: Helping or harming emergency management? via the National Disaster Preparedness Training Center (NDPTC)

PS: I hope you've enjoyed this newsletter! Creating it each weekend is a labour of love that I provide for free. If you've found my writing valuable, the best way to support it is by sharing it with others. Please click the share links below to spread the word with your friends and colleagues; it would mean so much to me. Thank you for reading!

Share on LinkedIn // Share on Facebook // Share on X (Twitter)

_{Parts of this newsletter were created using AI technology to draft content. In addition, all AI-generated images include a caption stating, 'This image was created using AI'. These changes were made in line with the transparency requirements of the EU AI law for AI-generated content. Some links in this newsletter may be affiliate links, meaning I earn a small commission if you click and make a purchase; however, I only promote tools and services that I have tested, use myself, or am convinced will make a positive difference.}