Opinion: Why we need to stay in control of artificial intelligence

Artificial intelligence (AI) has the potential to drive fundamental change in the monitoring, data interpretation and reporting on farms, enabling “big data” to bring productivity, health and environmental benefits to all livestock farming businesses.

But we are currently approaching a tipping point, because there are no plans for how the industry ensures the full benefits of AI are delivered in a controlled manner.

See also: How beef units could improve efficiency with camera tech

It would be naive to expect Defra to play any role in supervision on behalf of the wider industry. So we need a degree of industry self-regulation and farmer influence to derive the maximum value.

AI is here to stay and is developing fast. But who do you trust?

How do you know that the systems are going in the right direction for farmers, are addressing the core issues, and will make the optimum decisions? The industry needs a much closer involvement in how AI develops.

For many routine admin and scheduling tasks, AI should have an immediate benefit, removing mundane, repeatable chores and freeing time for other activities.

But how and when will systems be developed that can be trusted with more fundamental tasks?

For example, AI could use data to indicate a health issue. It might even advise a treatment. But can it replace stockmanship skills?

And what are the welfare and industry perception implications if using AI leads to unnecessary or missed treatments?

Engagement 

Without proactive and ongoing engagement from the industry, the risk of poorly researched and developed systems landing on farms is considerable, and the consequences potentially significant.

We know many AI systems are in development, but there is a lack of regulation or overall governance.

This is not to say that the process of AI development must be stringently regulated – innovation must be encouraged.

But a degree of industry-led regulation could deliver significant benefits and ensure that producers and the entire supply chain can derive value.

This needs to happen quickly, as once systems start being delivered widely, it will be impossible to apply controls retrospectively.

Statements

Many industry bodies are actively developing AI position statements, but the issue is wider than individual representative organisations.

The livestock farming industry will be best served by some form of industry-wide representation including producers, processors, the supply trade and vets. Such a body could perform several roles.

It could provide a point of contact between the industry, developers and investors, and help shape the development of systems.

It could help ensure the systems meet the real needs of the industry, while allowing developers to challenge the industry with new possibilities.

It could provide some form of quality assurance on the systems produced, perhaps introducing an industry standard.

This will allow producers to be more methodical in choosing the systems they want to employ, and help build trust in what is being developed.

It will put the industry more in control of its AI supported future.

AI will never replace farmers, but can be seen as a valuable co-pilot, supporting the running of the business.

The priority for the industry should be to ensure it has access to the best-developed and most trusted co-pilots.

The AI (Live) conference on 23 September in London will draw together people from across the livestock farming and AI development spectrum, providing a forum to air some of the issues facing the industry. Find out more at ailive.farm