This was according to a panel of industry experts who gathered at the event, F&A Next, to discuss what the future of AI and the food ecosystem will look like.
The panel, Exploring new horizons in food through artificial intelligence, was moderated by Kate Murray, investment manager and director at Nector Holdings. Panellists included Ioannis Athanasiadis, professor of artificial intelligence and data science at Wageningen University & Research; Alon Chen, CEO and co-founder of Tastewise; Olaf van der Veen, CEO and co-founder of Orbisk; Roland van der Vorst, head of innovation and CEO of Rabo Carbon Bank at Rabobank; and Jonathan Berte, founder of RoboVision.
What do the next five to 10 years hold for AI in food and ag-tech?
Chen believed AI will revolutionise how the food industry understands consumer intent and decisions. According to Chen, consumer behaviour will be more easily understood, and AI will make this data extremely clear and actionable.
He said: "In five years, it's going to be extremely clear to all of us what people eat and drink and why. This will help us transform, reduce food waste, increase well-being, [...] and I think putting the consumer in the centre [...] needs to be a clear mission to anyone who wants to stay in business.”
Van der Veen discussed operational improvement and how AI will enhance efficiency and sustainability in food services by handling repetitive and mundane tasks. He highlighted the potential of AI to act as a “copilot” for chefs and food service operators, allowing them to focus on delivering excellent products and services, while AI handles the mundane, repetitive, and routine tasks. He explained how AI feeds into the “inherent laziness” of food service operators and end users.
For Berte, the not-so-distant future will be all about efficiency and automation. Berte envisioned a future where AI and robotics will provide detailed insights, and optimise logistics, especially in produce markets. Berte explained how AI can help address the many challenges in managing varying quality and demand in produce: "With robotic platforms entering the scene, we will have a very detailed view of each and every tomato, each and every cucumber. [This] will lead to a very sophisticated segmentation of fruits and vegetables in the supermarket."
Berte cited Walmart’s future ability to use its forecasting models to predict a correlation between cherry tomatoes and a heat wave. This information would enable the company to then direct a pool of greenhouse suppliers to start harvesting before the heatwave, to supply that peak.
Barriers more likely to come from humans than technology
Murray raised the issue of the gap between AI development and practical adoption.
Athanasiadis explained that the primary barriers to AI adoption are human, not technological. He said that while younger generations adapt to AI faster, older generations struggle: “My generation learned how to Google with the right keywords. The younger generation is learning how to prompt AI, and that's a different skill."
Van der Vorst agreed with Athanasiadis, adding: “Where are the barriers? They are not on the technology side. They are on the human side, on the business side, the geopolitical side.” He said that when it comes to AI, the promise is that technology adapts to reality, not the other way around.
Chen outlined that barriers often come from within companies. He stated that while leadership often drives AI adoption, there is sometimes resistance from within teams.
He said: "At the very beginning, we [Tastewise] only worked with early adopters, [...] but what we are seeing today is organisations understanding they need to democratise the AI and democratise the data.” Nowadays, according to Chen, the consumer is moving so fast that if a company tries to outsource its next innovation to an agency, by the time it gets a response everything has already changed, and the company has lost the market. “If you are not getting the skills internally, you're losing the game," Chen added.
For Van der Veen, successful adoption is all in the simplicity of AI systems for end users. This approach helps reduce the resistance of AI and the fear associated with it. According to Van der Veen, users do not need to understand AI to benefit from it, and successful AI companies know this and hide the complexity of AI from users to make it more accessible. "We hide our AI. It's very hard to find on our website that we are an AI company. [...] because the end users do not care that they're using AI. [...] They just want a product that works,” he said.
Authorities need to regulate while still allowing space for innovation
Murray inquired about the role of governments, regulatory agencies, and institutions in facilitating AI adoption.
Chen explained how the European market needs to allow space for innovation to propagate while balancing this with privacy and security regulations. “In Europe, I think, just don't destroy it [AI],” he said. "The Privacy Act, right? Click to accept for cookies, was the biggest invention from Europe, and it actually gave a lot of power to the big corporations and killed so many startups."
Berte called for governments to build ecosystems that support AI innovation without over-regulating. He highlighted the importance of government involvement in creating favourable conditions for AI development. "We know that it's happening now, AI, and they [governments] have to feed the ecosystem, not only in a regulatory way," he said.
Van der Veen had only one word to say on the topic, “ethics”. He called for global standards to ensure AI is used responsibly: "The only thing I do expect from global authorities is ethics."
