Aquaculture North America

Features
Artificial intelligence 101


May 13, 2020
By Lalou Ramos
Credit: © pathdoc / Adobe Stock

When asked about his view on Artificial Intelligence (AI) Microsoft founder Bill Gates was quoted as saying, “The world hasn’t had that many technologies that are both promising and dangerous.”

Another tech giant, Elon Musk, refers to AI as “our biggest existential threat.” Yet despite such intimation of dystopian doom and gloom with AI-powered bots subverting humanity, various industry verticals— including aquaculture—continue to herald the benefits of AI and Machine Learning (ML) and the positive ways it disrupts the industry. Does AI pose any risks to aquaculture today? Or should we welcome the technology with open arms? Our in-house techie Lalou Ramos weighs in.

Artificial Intelligence or the use of computer algorithms to perform tasks that normally require human understanding isn’t new, but why are people talking about it as if it’s a recent discovery?

That is correct, Artificial Intelligence has been around for nearly 60 years but it’s only recently that we see tangible applications across different industry verticals including aquaculture. Actionable insights generated through Machine Learning are being used extensively to enhance processes and experiences that promise greater efficiency and profitability for fish farmers.

Mechanical and electronic devices such as automatic feeders have been in existence for years, why did it take decades to make such equipment “smarter”?

AI requires massive computing storage and processing speeds to store data, classify and label them, and get them to execute complex tasks quickly. Those were not yet available during the early days of AI. The availability of computers with greater processing power, the option to store data in cloud and the ability to push or access data in mobile devices all help to accelerate the cross-industry adoption of AI. These factors also enable machine learning, where computers learn and improve from experience without explicitly being programmed.

If machine learning is about collecting data and using it to help computers learn, what kind of data is being collected to create AI-based aquaculture applications?

Machine learning algorithms learn models from historical data. In the aquaculture context, data sets could come in the form of farm practices, yields, and environmental data sources. Each data is classified and labeled to ensure that they are accurate enough to reflect real vision of the world or market.

Should we be concerned about such data being compromised?

Data breaches are some of the unfortunate consequences of building a hyper-connected and digitized business environment. The good thing is there are encryption and data protection technologies that address different types of security threats. In addition, regulations on the storage or usage of data are constantly being reviewed and improved to secure data and protect data owners.

That said, certain level of openness and sharing of best practices and data among aquaculture industry players are necessary for the industry to move forward. For instance, the development of diagnostic platform for fish diseases using neural networks may require international collaboration and the gathering of data and information from volunteer farms around the world. The more data shared and gathered the more accurate and useful the outcome.

What are some of the promising AI-driven applications available in the market?

We see many new AI-based applications capitalizing on the ubiquity of mobile phones to make aquaculture platforms that are easy to deploy, use and maintain.

For instance, from Japan and Singapore came a smart feeder that can be controlled remotely via smartphones. This helps farmers optimize feeding schedule to reduce waste and increase profitability at the same time help them save time and maintain better work-life balance. In India, a company made a mobile application that helps shrimp farmers predict diseases and enhance water quality.

We see drones and robots that collect many different types of data ranging from water pH, salinity, dissolved oxygen levels, turbidity, pollutants which can help farmers make business decisions more accurately and profitably. There is also a company that deploys robo-fish to detect underwater pollution sources near fish farms.

Those apps are highly beneficial and not sinister at all. But going back to Elon Musk’s warning, how can we be sure that the aquaculture industry will not be run by robo-fish/autonomous weapons in the near future?

There are AI applications designed to automate tasks associated with human tacit knowledge or personal knowledge of behaviour and perceptions of people but the world has not turned a blind eye on the threats of AI and how it can potentially impact jobs or human safety. There will be watchdogs for AI. For instance, top universities including Stanford University, UC Berkeley, and MIT have established human-centered AI (HAI) research, which advances the idea that the next frontier of AI is not just technological but also humanistic and ethical. They believe that AI should enhance humans rather than replace them.


Print this page

Related




1 Comment » for Artificial intelligence 101

Leave a Reply

Your email address will not be published. Required fields are marked *

*