EU AI Act: Prepare now for major artificial intelligence law, says Rabobank analyst

23 Jan 2024 --- Food corporations are transforming their businesses through AI but should prepare now for the incoming EU AI Act, according to Julia Buech, a senior consumer foods analyst at Rabo Research Food & Agribusiness. The world’s first major AI regulation, which is expected to come into full force in 2026, will limit AI systems based on their risks to society.

In December 2023, the European Parliament reached a provisional agreement with the Council on the AI Act. The Parliament’s priority is to ensure AI systems used in the EU are safe, transparent, traceable, non-discriminatory and environmentally friendly. EU authorities also want to establish a technology-neutral, uniform definition for AI that could be applied to future AI systems.

“It definitely seems to be a bit of a lawless jungle out there at the moment when it comes to AI — that’s a widespread sentiment — and regulation is lagging behind,” Buech tells Food Ingredients First.

“However, regulation is underway within the EU. The EU AI Act aims to hold companies more accountable for what they’re doing with AI, taking a risk-based approach with different requirements for different risk categories — basically, the higher the risk, the stricter the rules, and the higher the fines.”

Non-compliance with the rules could lead to fines ranging from €35 million (∼US$38 million) or 7% of global turnover to €7.5 million (∼US$8.2 million) or 1.5% of turnover, depending on the infringement and size of the company.

“Companies need to prepare now, linking the data strategies and risk management more closely than before. The idea is to move toward a more secure and trustworthy AI environment that should hopefully benefit everybody,” says Buech.

AI elevates consumer food
In a recent report, Rabo Research identified several ways in which AI systems are elevating the food consumer sector despite the technology still being in its infancy. The core areas food companies are targeting are operational efficiency, customer experience and new product development.

“Companies are exploring AI to help with demand forecasting and more accurate inventory management. The sudden rise of generative AI also means that food companies can engage with consumers in novel and advanced ways — we see retailers offering hyper-personalized shopping advice to customers based on individual budgets, taste and cuisine preferences,” Buech tells us.

“Product innovation is also a key focus area — companies use AI to innovate around flavors, tastes and ingredients, and many of the bigger F&B players have partnered up or are in the process of partnering up with AI start-ups in these fields of tastes and ingredients to drive innovation.”

Longer term, Buech expects AI to play an important part in the transition to a more sustainable food system, including in the areas of more advanced alternative proteins, plant-based proteins, and functional nutrition.

“AI washing” and AI-free
AI systems also pose significant risks to the industry, including affordability issues, “AI washing” and the dangers of irresponsible use, which are issues the EU AI Act will work to address.

“‘AI washing’ — named after greenwashing — is a risk. For example, in product innovation, it can occur when companies claim that the offerings involve AI technology, but in reality, the connection to AI is minimal or not core to the product functionality,” explains Buech.

“The reason why companies could be doing this is to make products appear more cutting edge than they are — this can really damage a company’s reputation.”

The concern around “AI washing” could spur an AI-free countermovement.

“Watch out for AI-free become an interesting marketing strategy or marketing point of differentiation,” continues Buech.Buech shares insights on AI systems in the consumer food sector.

“There’s so much distrust with regards to AI potentially being a threat to humanity — you will naturally find a lot of consumers attracted to the ‘real thing.’”

The European Parliament wants to establish a technology-neutral, uniform definition for AI that could be applied to future AI systems. Currently, the EU AI Act uses the definition proposed by the OECD: “An AI system is a machine-based system that...infers from the input it receives how to generate outputs such as predictions, content, recommendations or decisions that can affect physical or virtual environments.”

Protecting small businesses
Another key concern is the cost exclusion barriers for smaller companies unable to risk investment in AI systems, which could lead to a greater power divide between big and small businesses.

“The costs of AI implementation vary significantly depending on factors such as the size of the organization, the scope and complexity of the AI solution and specific use cases. Large companies often have more resources and capital to invest in AI infrastructure, making them the likely initial winners in the looming AI era,” says Buech.

“Regardless of whether companies are large and powerful or small and specialized, we expect to see a deepening gap between technological leaders and laggards, which can lead to a widening imbalance of power.”

In the EU AI Act provisional agreement, MEPs made clear they want to ensure that businesses, especially SMEs, can develop AI solutions without undue pressure from industry giants controlling the value chain.

The agreement promotes so-called regulatory sandboxes and real-world testing, established by national authorities to develop and train innovative AI before placement on the market.

This week, the European Commission will reportedly push back on a US-led attempt to exempt the private sector from the international treaty on AI in a plenary meeting (January 23–26). 

European authorities also want the world-first treaty to align closely with the EU AI Act.

By Joshua Poole