AI is set to unwrap the packaging revolution 

AI is poised to provide smart solutions to packaging problems, helping to take businesses to the next level Credit: hxdbzxy via Shutterstock

Artificial intelligence (AI) shows significant potential to revolutionise the packaging industry, from product design solutions to cost- and waste-reducing efficient process engineering.   

According to Future Market Insights (FMI), a Pune, India-based market research provider, around 65% of global packaging organisations plan to invest in AI solutions in the foreseeable future.  

FMI predicts the value of AI in the packaging market to be just over USD2 billion in 2023, rising to over USD6 billion by 2023 – a CAGR (compound annual growth rate) of 11.2%.   

Major brands including UK-headquartered international alcoholic beverage giant Diageo and US-headquartered consumer brands Clinique and Johnson & Johnson are all, for instance, using AI to speed up packaging designs.   

Meanwhile, technology companies are working to deliver solutions that will integrate AI into the manufacturing process for packaging, rather than just the design phase.   

Toronto, Canada-based AI Materia has developed a trained AI platform to accelerate decision-making processes for better and faster packaging product development, by leveraging smart data infrastructure and advanced machine learning tools.

AI Materia’s platform works by integrating technical materials and processing data with established scientific insight into packaging materials production to provide actionable solutions for more efficient manufacturing.  

“[Using our AI technology] can result in 90 percent reduction in development time and cost for our customers,” Maryam Emami, CEO of AI Materia, told a 2023 Industry Open House and Manufacturing Forum at the McMaster Manufacturing Research Institute, at Hamilton, Ontario-based McMaster University, in Canada in June this year (2023).   

Emami also noted that “AI has become shorthand for any kind of cutting-edge technology,” when in her view the term should only be used for technologies that can replace human cognitive functions, such as perceiving, learning, interacting with the environment and problem solving.  

This is important, Emami said, for industrial users to fully understand and appreciate the potential of AI. Examples of this kind of 'cognitive replacement' technology, according to Emami, include machine learning, robotics, computer vision and language processing.   

Older types of AI, such as automation, have been integral to the packaging industry for decades, however the sector is still getting to grips with newer forms of technology, such as large language models (LLMs) like ChatGPT, according to Ismail Sutaria, chief analyst, packaging domain, at FMI. 

Speaking on a July (2023) webinar entitled 'The Power of AI in Packaging 2023', Sutaria said: “AI will become accepted [in the packaging industry] in the same way robotics is today, within the next three-to-four decades.”  

Sharina Perry, founder and inventor of Utopia Plastix, an Edmond, Oklahoma, USA-based company that develops plant-based alternatives to plastic packaging, said that “machine learning is the most applied AI technology in the packaging sector today.  

It can be used throughout the value chain to help companies with predictive maintenance and to improve their cybersecurity.”     

But Perry said that new iterations of AI into the packaging production process, such as LLP systems, could go further, making entire packaging supply chains much smarter.   

According to Sutaria, fast-moving consumer goods manufacturers are the leading adopters of AI in packaging, followed by retail, homeware and cosmetics companies.  

Sutaria said that these sectors are focusing on AI’s potential to optimise user experience, product integrity and managing inventory – with AI helping strike the right balance between sustainability and what consumers want, where previously brands relied on slow and expensive trial and error.   

However, Sutaria thinks that AI’s much-touted potential to significantly reduce costs is still some way off: “Many organisations are still at the experimentation stage when it comes to AI. Unless and until the scale of adoption is at mass level, bringing down the cost curve is going to be a challenge,” he said.  

Barriers to mass adoption include the high cost of technology and concerns about exposure to security vulnerabilities, Sutaria noted.  

Philadelphia, US-based digital marketing agency WebFX estimates that a single custom-designed AI solution could cost a business anywhere between USD6,000 and USD300,000 in 2023, with ongoing maintenance and consultancy costs on top, while third party ‘off the shelf’ solutions could cost anywhere from nothing to USD40,000 per year.   

“We need to start with semi-automation and then shift gears towards more AI-powered processes. Brands need to keep an eye on the switching costs, as AI will require high investment to start with. So [the industry needs] to work on cost reduction,” Sutaria added.  

Perry argued that focusing on AI’s potential to cut costs by replacing humans in certain tasks is the wrong way to think about the technology, from a sustainability point of view. 

“In my opinion, if we take that approach, we have missed the mark on sustainability. We need to learn how to incorporate technology without adversely affecting human life,” she said, noting that AI can be deployed across every node of the value chain to improve packaging sustainability.  

“The core focus of AI when we’re talking about packaging has been in manufacturing. So now we must rethink AI usage, from manufacturing, to consumption, to waste and to the resources [that go into making packaging],” she added.  

But Perry stressed that AI is not a clean panacea for more sustainable packaging production. “One thing many people don’t realise is that AI programmes consume a lot of water. The cloud-based data processing centres [AI relies on] face high water costs due to power and cooling [requirements], so AI has a hidden water footprint,” she added.  

Other areas where AI currently falls when it comes to delivering improvements for packaging is in its solutions for product testing and recyclability certification, according to Krzysztof Krajewski, chief sustainability and innovation officer at Italy-headquartered RDM Group, a leading producer of recycled packaging materials.    

“Algorithms [are being used to] design the scale of recyclability of products. But simple AI models are not delivering,” Krajewski explained, noting that “AI-based certifications for the recyclability of products do not stand up to results of real-world testing.”    

But while he is sceptical about AI’s certification capabilities, Krajewski does believe it can be deployed effectively in the manufacturing process: “At the moment, most packaging ends up in landfill or incineration at best. If we improve our processes with AI, we can end up with much cleaner waste, which isn’t really waste anymore because we [will be able to] recycle it.”  

A challenge Krajewski sees to achieving AI’s potential in the packaging sector is that many companies do not know how to apply the technology. 

“It’s about looking at what problems you want to solve first. That usually means looking at the changing regulatory landscape. Identifying the problem is the first step, then you need to look at whether you have the mass data for the machine learning to learn from and make decisions. If not, you need to think about how to collect this data,” he said.