Like all industries, the reach of artificial intelligence (AI) is expanding throughout agribusiness. There are the commonly used “chat bots”, alongside primary production specific technology such as cattle monitoring equipment and the ever-growing world of automated farming machinery.
At the heart of this expansion is data. But where does the information fed into AI software go? Who owns it? And who controls it?
The answers are not straightforward, with “terms of use” varying across AI service providers and differences between free and paid AI tools. Such answers are also hidden in the “fine print” and often missed in the process of subscribing to a provider.
Ownership vs. Use
In Australia, there is no legal ownership claim over “data”. For that reason, the key issue is not who “owns” the data but rather how an AI provider might be entitled to “use” that data. Most policies allow for the provider to use data inputs for training purposes, while some providers even allow for disclosure to third parties as “sub processers” of data.
In more extreme circumstances, a provider might use the data to influence commercial decision-making. This was the case revealed by the recent International Panel of Experts’ Report on Sustainable Food Systems, where top tech giants like Google, Amazon and Microsoft were found to be using data to “play with the food system”. By harnessing information collected from satellite and drone sensors, which monitor climate conditions and soil health, they’ve been able to influence farmers to grow seeds sold only by multinational agricultural companies.
With this promotion of international products over the locally adapted crops grown by farmers for generations, there are widespread concerns about food security given the potential for farmers to be “locked in” to a globalised market.
Security of information
If an AI provider’s terms allow it to “use” data for certain purposes, it will be stored on the provider’s servers (and potentially third-party servers). This prompts the question – how secure is the information? Commercially sensitive operational or financial data that is fed into AI software is potentially being exposed to an AI provider’s security breach.
Weaknesses in AI providers’ databases have already been exposed:
- in late 2025, ChatGPT’s third party “analytics” provider was hacked which resulted in usernames and email addresses being leaked; and
- on a more serious scale, in February 2026, an independent security researcher was able to access 300 million messages from over 25 million users in the chatbot, “Chat & Ask AI”.
The way forward
Such worries may explain the relatively cautious uptake of AI technologies across the Australian agribusiness sector. To address concerns, the National Farmer’s Federation has developed the “Australian Farm Data Code”. However, the Code is not legally binding and applies only where providers choose to participate. There seems to be some way to go before there is regulation and transparency across AI providers.
The spread of AI technology seems inevitable though we suggest that producers should take a considered approach to adoption, including by taking the following steps:
- avoid having sensitive data fed into AI chatbots, particularly when using free versions of software;
- where possible, anonymise data to ensure that if such data is caught in a security breach, it cannot be traced back to you as a user; and
- use reputable providers rather than smaller, unverified operators.
Thynne + Macartney’s Agriculture team can assist primary production businesses with reviewing the terms and conditions of an AI service provider and advising on the legal and commercial risks associated with adopting these technologies.
About the Author
