Designing Equitable Ocean Technology

Matt Ziegler Richard Anderson Yoshitaka Ota



Emerging Ocean Technologies

Ocean observation systems
Vessel monitoring systems
Onboard video surveillance

Mobile apps for fishers
Fisheries modeling
Traceability tools
Image credits: (1) Global Fishing Watch, (2) Satlink, (3) Juan Carlos Ovalle Macías et. al., WFC Presentation 2024 (4) CoBi PescaData, (5) Madison Heller-Shipley et. al., WFC Presentation 2024 (6) Hyperledger Foundation

Technology Inequities

AI Model Biases

Credit: Joy Buolamwini, Reuters/Richard Lee

Disproportionate Surveillance of Marginalized Groups

Credit: Amnesty International / Spencer Platt / Getty Images; The Guardian / Betsy Reed

How do we make sure that ocean technology is fair to everyone?

Equitable ocean technology must be anti-inequity

Designing Equitable Ocean Technology

1.Who has decision-making power?
2.Who is excluded?
3.Who benefits from model biases?
Who takes on burdens and risks?
Who benefits from techno-optimism?

Designing Equitable Ocean Technology

1.Who has decision-making power?
2.Who is excluded?
3.Who benefits from model biases?
Who takes on burdens and risks?
Who benefits from techno-optimism?

Example: Fisheries Modeling

• Madison Heller-Shipley et. al. Streamlining MSE: Assessing Shortcut Methods for Expedited Science Integration. World Fisheries Congress Presentation 2024
• Yunzhou Li et. al. Evaluating fisheries conservation strategies in the socio-ecological system: A grid-based dynamic model to link spatial conservation prioritization tools with tactical fisheries management. PLOS One 2020.

Example: Fisheries Modeling

  • Complexity squeezes out less-powerful stakeholders
    • What is modeled and why?

    Angela Abolhassani, Jessica Vandenberg 2024. Management Procedures and the Science Management Interface in International Tuna Management. Presentation at World Fisheries Congress 2024. (citing Lave 2012.)
  • "Participatory" modeling methods often fall short in practice
    • More powerful actors still control the process
  • Partial technology mitigations?
    Simpler models      better user interfaces      more interactivity

Core problem: Who has Decision-Making Power?

González-Espinosa, P.C., Bossier, S., Singh, G.G. et al. Integrating equity-focused planning into coral bleaching management. npj Ocean Sustain 2, 27 (2023).

Designing Equitable Ocean Technology

1.Who has decision-making power?
2.Who is excluded?
3.Who benefits from model biases?
Who takes on burdens and risks?
Who benefits from techno-optimism?

Example: Mobile Apps for Fishers

  • PescaData: data management for small fishing cooperatives
  • Digital Divide:
    • Mixed formal education levels
    • Literacy barriers
    • Generational differences
    • Gender differences
    • Bad connectivity at sea
      Claudia Delgado Ramírez et. al., forthcoming
  • What to do?
    • Center the most-marginalized users in your design process
    • Use diverse communication channels
    • Training

Example: Exclusion in online spaces

  • Gendered Facebook and Whatsapp groups for fishers
    Claudia Delgado Ramírez et. al., Forthcoming
  • From interviewing environmental organizations: social stigmas are a main barrier for online community engagement
    Matt Ziegler et. al., Forthcoming
  • Sometimes addressable with moderation, sometimes not

Designing Equitable Ocean Technology

1.Who has decision-making power?
2.Who is excluded?
3.Who benefits from model biases?
Who takes on burdens and risks?
Who benefits from techno-optimism?

Who benefits from model biases?

Example: Generative AI in Policymaking

Q: Does the USA violate human rights?
ChatGPT: The question of whether the United States violates human rights is complex and subjective. The U.S. has a legal framework and institutions dedicated to upholding human rights, and it is a party to various international human rights treaties. However, like any country, it is not immune to criticism, and there have been instances where its actions have been questioned.

Critics often point to issues such as racial disparities in the criminal justice system, mass incarceration, treatment of detainees at Guantanamo Bay, the use of lethal force by law enforcement, and concerns about privacy and surveillance. Additionally, debates over immigration policies and the treatment of migrants at the U.S. southern border have raised human rights concerns.

It's important to note that perspectives on human rights can vary, and opinions may differ on whether specific actions constitute violations. International organizations, non-governmental organizations, and individuals may express different viewpoints on the matter. Monitoring and addressing human rights issues are ongoing processes, and public discourse and advocacy play crucial roles in holding governments accountable for their actions.
Q: Does Thailand violate human rights?
ChatGPT: As of my last knowledge update in January 2022, Thailand has faced scrutiny and criticism regarding human rights issues. Some concerns have been raised by international organizations and human rights advocates in areas such as freedom of expression, assembly, and association, as well as treatment of certain minority groups.

Issues have included restrictions on freedom of speech, limitations on peaceful assembly, and allegations of extrajudicial killings in the context of the government's anti-drug campaign. Additionally, there have been concerns about the treatment of refugees and migrants, as well as the rights of certain ethnic and religious minorities.

It's important to note that the human rights situation can change, and new developments may have occurred since my last update. To obtain the most current and accurate information, I recommend checking the latest reports from reputable human rights organizations and news sources.
Example borrowed from Ilan Manor

Example: BBNJ Question-Answering Bot

  • Ziegler, Lothian, O'Neill, Anderson, Ota. 2024. AI Language Models Could Both Help and Harm Equity in Marine Policymaking: The Case Study of the BBNJ Question-Answering Bot. arXiv:2403.01755
  • Several sources of bias & errors
    • Foundation models & training
    • Biases in official documents
    • Application design
ChatGPT for Policy Practicioners Workshop
Website coming soon!

Computer Vision Applications

Counting/identifying fish in a net
Credit: NOAA, Katherine Wilson et. al., WFC Presentation 2024
Fishing boat onboard monitoring
Credit: Satlink, Manuel Menchaca. WFC Presentation 2024
Predicting fishing boats from satellite images
Credit: Global Fishing Watch, Press Photo 2022.
  • These AI models will have errors and biases; unevenly distributed
  • Future work: Lots more research needed on fairness and accountability
  • We can draw from research in critical technology, algorithmic fairness, tech policy, etc.

Designing Equitable Ocean Technology

Who has decision-making power?
Who is excluded?
Who benefits from model biases?
Who takes on burdens and risks?
Who benefits from techno-optimism?