08
Sun, Dec

AI: when’s it coming to this sector?

There’s plenty of talk about artificial intelligence and how it’s going to impact life as we know it. But to many it’s still a nebulous concept. So what’s the situation when it comes to its deployment in the large-format print sector?

Question 1: Where, as far as you are aware, is AI already being used in the digital print sector?

Dr. Nandini Chakravorti, Technology manager, data and information systems, The Manufacturing Technology Centre (MTC): 

Each day, around 5m devices connect up with each other, with the internet, or with both. There are around 6.4bn data-communicating objects in the world today. And by 2020, this number is forecast to explode to around 20bn. Emerging disruptive technologies such as Artificial intelligence (AI), robotics, and the Internet of Things (IoT) have the potential to have a huge impact in industries and society. In a 2017 report, the World Economic Forum identified a $100 trillion opportunity for both industry and society through the adoption of digital technologies.

The ultimate goal of AI, is to have machines and systems capable of performing tasks and cognitive functions that can only currently be performed by human intelligence. In order to perform these tasks, machines must be able to learn these capabilities automatically instead of having each of them be explicitly programmed end-to-end. Today AI is being used in a number of different sectors such as financial services (for fraud detections, speech recognition for automated customer service interactions), health care (computer vision to automate the analysis of medical images) and media (Netflix uses machine learning to predict which movies a customer will like). In manufacturing in particular, AI has multiple applications such as the potential to predict maintenance of assets to minimise downtime, identify product defects more reliably than the human inspection to reference but a few.


Mark Stephenson, Product manager, digital printing and press systems, Fujifilm Graphic Systems Europe:

Low-level AI has been with us for decades. Anything that is automated, robotic or that replaces a human decision or action qualifies as AI. A press that monitors itself for faults, software that controls colour or page layout or even simple mail-merge functions can be described as AI.


Gudrun Bonte, Vice president, product management, SAi:

In some areas of the industry, AI has already started to make in-roads, and more and more PSPs are beginning to notice the potential that implementing AI could soon have on their businesses. 

SAi is already exploring the potential and has launched a collaborative three-year research project with Oxford University’s Department of Engineering Science.

While AI as a whole is still in its early stages, we are starting to see which areas of the digital print sector it be effected. For instance, new technology is helping to reduce waste in printing, with smart software that is capable of calculating the best layout for documents to make the best use of available space. AI could save businesses time, money and even improving their global footprint. Similarly, AI could aid inspection and quality control systems, with the potential to improve speed and accuracy thanks to intelligent algorithms. 

Question 2: Are there particular areas where it will be deployed first – and why?

Dr. Nandini Chakravorti, Technology manager, data and information systems, The Manufacturing Technology Centre (MTC): 

According to the ‘Made Smarter Review 2017’, the UK has the strongest AI and machine learning market in Europe, with over 200 SMEs in the field (compared to just 81 in Germany and 50 in both the Nordics and France). The UK has a tremendous platform to capitalise on disruptive technologies like AI. 

AI has huge potential in a number of use cases within manufacturing such as, predictive maintenance, quality management and enabling faster more reliable design, to name a few. Within the digital print industry context, an example might be intelligent printers which through the advent of IoT-based sensors, could be instrumented and become more data rich, enabling predictive algorithms to be used to monitor the health of printers, or even generate service requests before a printer fails. An AI augmented technician can then recommend software updates or other necessary adjustments with minimum disruption to production. 

However, the adoption of AI in industries has been slow due to a number of reasons including: lack of leadership knowledge and awareness, lack of support for start-ups, and a lack of clarity on the return on investment. Developing an AI strategy with clearly defined return on investment, finding individuals with the appropriate skill sets, clear ownership and strategic commitment to AI on the part of leaders are some of the ways forward to ensure AI is deployed to get benefits.


Mark Stephenson, Product manager, digital printing and press systems, Fujifilm Graphic Systems Europe:

It’s already started, as described in the answer above, but there is certainly more to come. PwC define three waves of AI. These are automation - Wave 1 (to early 2020s): algorithmic, Wave 2 (to late 2020s): augmentation, and Wave 3 (to mid-2030s): autonomy

If we’ve already ticked the box for the first wave, then the second wave - augmentation, where AI starts to work alongside humans - is already well under way. 

In the world of digital print software analytics and dashboards, such as Fujifilm getFIT, analysis tools are advising customers of the most cost-effective or time-efficient print technologies to use for a given batch of jobs. Also, cloud-based colour management tools like Fujifilm’s ColorPath can predict potential colour consistency issues across a fleet of devices or a cohort of operators.


Gudrun Bonte, Vice president, product management, SAi:

If our own project with Oxford University runs successfully over the coming three years, it is possible that the development of an AI-assisted software solution could take place and become the first of its kind within the sign and display industry. 

One of the key ideas envisaged with this project is a software with the ability to inspire designers as they work, by providing suggestions such as best practice fonts, images and layouts. Being able to take advantage of ideas that an AI tool is trained to supply will really boost efficiency and throughput. At this stage, even toying with what might be possible later down the line is creating a real buzz within the industry, so I can’t wait to see how the print industry evolves when AI becomes more and more prominent.

Question 3: Do you think people running digital print companies understand the potential of AI to their business?

Dr. Nandini Chakravorti, Technology manager, data and information systems, The Manufacturing Technology Centre (MTC): 

The application of AI is a paradigm shift from inflexible solutions to adaptive self-learning solutions based on huge amounts of data and intelligent algorithms. The application of AI in industry offers £198.7bn value at stake to the UK economy between 2017 and 2027 (‘Made Smarter Review’). A 2016 study by global management consultancy firm Accenture suggests that the productivity enhancing impact of AI can add £650bn GVA to the UK, with a productivity level 25% higher than it would be otherwise (Accenture, ‘Why Artificial Intelligence is the Future of Growth, 2016’). 

Though AI and machine learning have been around since the 1950s, the technology’s ability to live up to its expectations is becoming possible due to the increase in high-performance computing and more widespread connectivity, thus making more data available. For example, the data captured in the printing process can be used to create intelligent algorithms to make real-time, in-process adjustments such as paper alignment and image quality. This can potentially lead to an optimal printing outcome without human intervention.

Printing companies have a significant opportunity to take advantage of existing software and equipment to inject AI in their businesses. AI can give print companies an opportunity to created targeted print catalogues for users based on their online viewing pattern.


Mark Stephenson, Product manager, digital printing and press systems, Fujifilm Graphic Systems Europe:

It’s a mixed bag. Some are grasping the developments and seeing the benefits while others will see AI as a threat. But the print industry is, in general, very good at adapting and integrating new technologies. Those who ignore the benefits will unfortunately be left behind. The survivors will be those who engage with their suppliers and partners to discover what value they can bring to the further adoption of AI in their business.


Gudrun Bonte, Vice president, product management, SAi:

 

It’s probably still a bit early for most people, so there’s clearly an education process required in order to ensure digital print companies are aware of its capabilities. I personally believe that any technology that presents an enabler for digital print companies to offer something extra to customers or enhance the way in which they work with customers, is worth exploring.

Question 4: By 2030, do you think AI will be making operational decisions within digital print companies?

Dr. Nandini Chakravorti, Technology manager, data and information systems, The Manufacturing Technology Centre (MTC):

The first companies employing AI systems across the board will gain competitive advantage. AI can be used to improve business performance in areas including predictive maintenance, where deep learning’s ability to analyse large amounts of high-dimensional data from audio and images can effectively detect anomalies in factory assembly lines.

However, it is a leap of faith to think that by 2030, humans will have learnt to use AI in a responsible manner, and we will have also learned to ensure that AI is regulated appropriately.  The ‘black-box’ complexity of AI techniques poses a challenge of explaining the outcome. It is of particular importance to generate trust and reliability of these techniques. We will also live in an improved world with potentially more use of intelligent robotics and automation in industrial processes. So, with this trend we will have a much bigger debate about up-skilling and re-training the workforce.


Mark Stephenson, Product manager, digital printing and press systems, Fujifilm Graphic Systems Europe:

This is the nirvana of the third wave, autonomy, where we have systems that learn and, rather than just follow rules, begin to make new, more efficient or appropriate rules. Systems that constantly improve and develop to become greater than their original brief or concept and make decisions on their own. However, we will always need humans to keep things on track. There are many stories of chat bots that develop racist tendencies or recruitment systems that pick men over women when left to learn from data alone. I’d like to answer this question more comprehensively but my task management system is reminding me that I have more pressing matters that require my attention.


Gudrun Bonte, Vice president, product management, SAi:

As the technology continues to evolve, I’m sure that AI has the potential to make operational decisions within digital print companies. Our hope is to offer an AI solution that aids the design process. As with any technology, all I can say is watch this space.