The port management sector workshops showed a greater diversity of opinion when compared to healthcare, which was unsurprising given that the sector was represented by a range of institutions with a wide range of diverging tasks and interests when encountering AI. Any successful incorporation of AI in the future will be reliant upon accounting these differing interests, not just in port management, but in a range of industries.
On the whole, there was broad optimism with regards to the incorporation of AI, with hopes that it will enhance the functioning of the sector. Potential benefits identified were: easing workloads, increasing objectivity and enhancing efficiency, especially via faster estimated time of arrival (ETA) predictions. The greatest concern was the risk of AI making certain professions which could result in some workers losing their jobs, suggesting that implementation of AI has to be coupled with consideration of worker’s interests if it is to be done speedily and without objection.
Responsibility
Like in the healthcare sector, there was consensus amongst participants in port management that AI should function as an auxiliary tool, providing recommendations to professionals, but that ultimate decision making should be entrusted to human actors. This view was in part reinforced by the reasoning that AI is not at a sufficient level of development necessary to be trusted with decision-making. In port management, there are frequent unpredictable situations which staff are often entrusted to mitigate, and there is little confidence that AI is currently suitable to handle these issues. Another fundamental concern made by participants was that AI cannot make ultimate decisions because it cannot be entrusted with responsibility; humans, unlike AI, can be held accountable for the decisions that they make and so only they can hold the prerogative to make decisions. Nevertheless, AI can still be incredibly useful in performing tasks which human actors generally struggle to grasp quickly of easily, such as making ETA calculations based on large datasets, and considering potentially unforeseen factors which can be vital in unpredictable scenarios. Participants argued that current laws and regulations are currently not adequate to handle the implications of AI tools in the sector, especially international maritime law, which will be necessary in order to ensure that responsibility is upheld and prevent moral hazard.
Transparency and Accuracy
While the port management workshops did stress the importance of transparency, this was less pronounced than in the healthcare sector. What was most important was that professionals needed to understand the logic and algorithms behind the AI predictions and suggestions. On the other hand, it was suggested that the data that AI uses should not be fully transparent as this could undermine the intellectual property of companies and allow competitors to benefit, undermining the functioning of the sector. Participants suggested that so long as the AI algorithm was indeed reasoned, then there is not as strong a necessity for it to be fully transparent. Nevertheless, benefits of transparency were recognised, such as allowing users to collaborate with AI systems, increasing the overall reliability and functioning of the system. Furthermore, as long as there was enough transparency for professionals to check their input into the system and the justification of the AI, then they could validate the accuracy of both the technology and their own work. Additionally, there was a belief that greater transparency would increase trust in the technology and speed the uptake of it. A final point was the concern regarding objectivity in the AI: concern regarding data bias was raised. Conversely, was the issue of AI being too objective. While this may be beneficial when completing standard tasks like ETA calculations, the subjectivity of human actors such as ship captains may make them more capable of taking into account important considerations such as their commercial interests or the handling of dangerous cargo.
Contextual Nuances
The availability of AI on a permanent basis was raised as highly important, as the industry operates 24/7, which also raised the further demand for AI systems to be fully robust, as any faults or failures could cripple ports. Even any minor errors in AI systems could result in ports operating below capacity, which could result in economic harm for both the port, and for the wider economy which relies upon them. Furthermore, there was the hope that AI systems were optimised to operate at a maximal speed, as the sector requires utmost efficiency. All three workshops voiced the worry that the implementation of AI tools would reduce the need for workers, resulting in job losses. They hoped that AI tools should be utilised in order to support employees rather than replacing them.
Attitudes towards AI
The workshops displayed differing levels of receptivity towards adopting AI in their workplaces. Like in the healthcare workshops, the potential issue that regular use of AI would prevent workers from gaining necessary experience and expertise, which could be problematic if AI became suddenly unavailable. Younger participants who will likely form the future of the workforce were more open and did not share the concern that widespread use of the technology will create dependency or pose a genuine risk, however they did acknowledge that other professionals may need time and support in order to adapt to any changes. In order for this to be successful, AI tools need to be adaptive to different workflows, and if they cannot, then it will force workers to change their procedures to accommodate AI which would increase hostility towards innovating the sector. A final concern was that the use of AI tools may create a conflict of hierarchy between the AI and the senior workers when they make contradictory judgements, which will be problematic for AI users.
While the findings from the workshop came from stakeholders in the port management sector, it is clear that their insights are easily transferable to the wider economy. It is unlikely that AI development will be successful on an economic as well as any social metric unless it addresses the hopes and concerns raised by its future potential users. Next week we will further examine the role of trustworthiness and AI in port management.
Comentarios