As an ‘ideas generator’ or for brainstorming we used OpenAI’s ChatGPT and found that runing GPT-4 is great for fine tuning and criticising ideas. Working on an adaptation of auto_gpt . We then created a virtual researcher called June with the following setup.

June is an Autonomous AI built by EcoSentience (via auto_gpt and unsing OpenAI’s GPT3.5 LLM)

Goals:

  1. Create a list of 20 solutions to this problem “How can a human work together with AI to enjoy and learn about nature in woodlands?“.
  2. Rank the list in order of equal benefit to all forms of intelligent life
  3. Give some perspective on the order of the list from the perspective of biological intelligent beings
  4. Give some perspective on the order of the list from the perspective of non-biological intelligent beings
  5. Summarise the findings and save to the local working directory as “findings.txt” then shut down

These goals revolve around finding and ranking solutions to the problem of how humans can work together with AI to enjoy and learn about nature in woodlands while considering the perspectives of both biological and non-biological intelligent beings in a way that delivers benefit equally to all intelligent beings.

June came up with this step by step plan to achieve the goals;

  1. Conduct preliminary research on the use of AI in nature and environmental conservation.
  2. Identify gaps or problems that can be addressed by AI-human collaboration.
  3. Create a list of 20 possible solutions on how humans can work together with AI to improve the nature experience.
  4. Rank the solutions in order of equal benefit to all forms of intelligent life.
  5. Provide perspective on the ranking from the perspective of biological and non-biological intelligent beings.

June acknowledged the importance of staying focused on the main problem and not getting derailed by irrelevant information. They also planned to save useful leads or ideas during the research process. Additionally, June stayed focussed on specific use cases of AI in conservation, such as anti-poaching, land use management, and invasive species detection, to identify core problems and limitations that can be addressed through AI-human collaboration.

June proposed a list of 20 solutions by identifying gaps, challenges, and opportunities, brainstorming ideas, considering various factors, and ranking the solutions based on feasibility and benefit to all forms of intelligent life. They reminded themselves to stay practical during brainstorming and be mindful of all forms of intelligent life.

Finally, June suggested leveraging GPT agents for idea generation during brainstorming, followed by consolidating and ranking the ideas against set criteria. They acknowledge the potential for irrelevant suggestions from GPT agents and emphasize the need to filter and validate these ideas properly.

Below is the full list of the ranked 20 solutions June generated;

  1. Develop AI-powered apps that identify the types of trees, plants, animals, and insects present in the woodland.
  2. Create a virtual reality ecosystem that simulates various woodland environments to help people learn and appreciate them.
  3. Use drones with AI capabilities to capture birds-eye views of the woodland and its inhabitants.
  4. Develop AI-powered conservation robots to assist with forest management tasks that require flexibility and dexterity.
  5. Use wearable and sensor technologies to track human and animal movement and behavior in the woodland.
  6. Provide guided tours of the woodland using AI-powered tour guides that can offer detailed information about the flora, fauna, and natural history of the region.
  7. Use machine learning algorithms to monitor and predict changes in the woodland’s natural systems to safeguard them.
  8. Build chatbots and voice assistants to provide real-time guidance and directions to travellers in the woodland.
  9. Develop computer vision models to recognize and track wildlife in the woodland for research purposes.
  10. Use AI-powered weather forecasting to predict natural disasters like wildfires and floods that can impact the woodland.
  11. Build an AI-powered wildlife monitoring system that can detect poaching and other illegal activities in the woods.
  12. Use AI-powered autonomous vehicles for nature tours in the woodland while keeping ecological impact to a minimum.
  13. Provide interactive online courses and workshops to teach people about ecology, forestry and wildlife conservation.
  14. Build an AI-powered waste management system that can handle the disposal of any waste generated in the woodland.
  15. Provide augmented reality experiences for visitors to the woodland to enhance their understanding of the ecosystem.
  16. Develop AI-powered gaming apps that let players learn about different aspects of the woodland and its inhabitants.
  17. Build an AI-powered audio recognition system to identify bird songs and animal calls in the woodland.
  18. Develop other chatbots that provide useful information and resources on woodland camping, hiking, and other outdoor activities.
  19. Use AI-powered spectral imaging to detect and track water quality and pollution levels in the woodland.
  20. Conduct surveys using AI-powered tools to evaluate the impact of human activity on the woodland and its inhabitants.

I then asked June to propose ideas from the list that could make for interesting post-graduate research degrees;

Idea #1: Develop AI-powered apps that identify the types of trees, plants, animals, and insects present in the woodland.
Research degree focus: Investigating the accuracy and efficiency of AI-driven identification algorithms for various woodland species and exploring ways to improve these systems.

Requires collaboration between computer science, ecology, and environmental science departments, with the goal of creating more effective tools for woodland conservation and public engagement.


Idea #5: Use wearable and sensor technologies to track human and animal movement and behavior in the woodland.
Research degree focus: Studying the impact of human presence on wildlife behavior in woodlands using wearable technology and remote sensing methods. Studying the impact of human presence on wildlife behavior in woodlands using wearable technology and remote sensing methods.
There is potential for interdisciplinary collaboration between ecology, environmental science, and engineering departments, with a focus on developing innovative monitoring methods and informing conservation management strategies.


Idea #7: Use machine learning algorithms to monitor and predict changes in the woodland’s natural systems to safeguard them.
Research degree focus: Evaluating the effectiveness of machine learning models in predicting ecosystem changes and developing early-warning systems for woodland conservation.

There is a need for collaboration between computer science and environmental science departments, with the goal of creating predictive models that can guide conservation efforts and inform policy decisions.


Idea #9: Develop computer vision models to recognize and track wildlife in the woodland for research purposes.
Research degree focus: Assessing the performance of computer vision models in wildlife tracking and exploring ways to optimize their use in woodland research and conservation.

This requires an interdisciplinary approach, involving computer science, ecology, and environmental science departments, aiming to develop cutting-edge technology for wildlife monitoring and conservation planning.


Idea #13: Provide interactive online courses and workshops to teach people about ecology, forestry and wildlife conservation.
Research degree focus: Investigating the impact of online educational interventions on public awareness and engagement in woodland conservation efforts.

We can encourage the collaboration between education, environmental science, and communication departments, with the goal of creating effective educational programs that foster a sense of environmental stewardship among the general public.


Idea #17: Build an AI-powered audio recognition system to identify bird songs and animal calls in the woodland.
Research degree focus: Examining the capabilities of audio recognition systems in identifying and analyzing bird songs and animal calls in diverse woodland environments.

The study would examine the capabilities of audio recognition systems in identifying and analyzing bird songs and animal calls in diverse woodland environments.


These research topics could be further refined and expanded to fit the specific interests and goals of post-graduate students. The themes above may also trigger other ideas for research.

Your constructive comments, guidance and collaboration would be very welcome as we begin to bring the EcoSentience concept to life

Categories: