AI has the potential to deliver significant benefits to the maritime industry, but several challenges must be overcome to fully unlock its capabilities. In the next few articles in this series of twelve, we explore some of these areas in greater detail.
Some of these challenges include:
- Data suitability and maturity
- Data consent
- Trust, transparency, and user acceptance
- Regulation
- Ethical concerns
The Data Consent Crisis
General-purpose and multi-modal AI systems are built on large quantities of public data, but today there is an emerging crisis in data consent. This means that there is a growing risk that AI platforms will face restrictions in getting access to the data they require to learn, develop, and make increasingly advanced decisions.
A recent study by MIT, which looked at 14,000 web domains, found that online platform owners are taking steps to prevent their data from being used by others. Across three common data sets, C4, RefinedWeb, and Dolma, 5% of all the data and 25% of data from critical domains have been restricted. Companies like OpenAI and Microsoft are now facing legal ramifications for copyright infringement. Twitter and at least eight US newspapers have announced plans to sue OpenAI and Microsoft for taking content from their news articles without permission or payment in order to train their AI chatbots.
The other concern for those that develop or own large datasets and share them with third parties, is the risk of their data being used to train a model their competitor might have access to. This could increase the importance of exclusively licensed datasets and proprietary data.
In the shipping industry, data lakes are increasingly being used by owners and operators to manage and increase the value of their data. A data lake houses clean and verified historical and live data, which can be accessed by algorithms running prediction programmes. This allows shipping companies to store, explore and search their data for insights. However, what happens when a large ship owner allows an AI model access to its data lake and this same model is used by a competitor? The AI model could have taken information from the data lake, used it to learn and develop, and provided even better outputs to another shipowner.
A data lake could cost thousands or even millions, depending on its features and complexity. Securing data lakes and implementing safeguards to manage data securely, efficiently, and in compliance with regulations are crucial to prevent a shipowner from losing their competitive edge.
To learn more about the challenges and opportunities in AI discussed in this article, read our thought leadership report, produced in collaboration with Lloyd’s Register titled, “Beyond the Horizon, Opportunities and Obstacles in the Maritime AI Boom” This report offers an in-depth analysis of the current state of operational AI in the shipping industry, shedding light on the latest trends, groundbreaking developments, and successful implementations. You can download a copy of the report here.
