Artificial intelligence

AI is still in its infancy but has huge potential for print

Pattern generated by Prinfab AI using a prompt
Pattern generated by Prinfab AI using a prompt

As with any new or emerging technology that is being talked up as the second coming, it can be wise to proceed with caution in its early days, at least until its potential is somewhat proven.

But while it’s impossible to tell at this stage how transformative artificial intelligence (AI) will be, both for print and on a general level, many early adopters are warning those that are ignoring AI, or are even fearful of it, to get a grip on its potential sooner rather than later to help their businesses thrive into the future.

“This technology has always been seen as science fiction but now all of a sudden it’s here, so I feel like, for a lot of people, that’s a lot to process,” says Keith Haas, solutions analyst at Keypoint Intelligence.

“A lot of the AI tools are new and people might not know what’s available, so I would encourage everyone in the industry to educate themselves because this is something that will affect everyone.”

In April, the UK government said it would invest £100m in start-up funding for a taskforce to develop safe and reliable AI foundation models.

These, which include large language models like ChatGPT and Google Bard, are generalist AI systems that have been trained on vast data sets, and can be used for a wide range of tasks across the economy.

Capable of tracking down and piecing together relevant information for specific queries, these models are also capable of generating original content that closely mimics anything a qualified human might be able to produce.

The government hopes that the development of sophisticated AI technology could be of major benefit to public services.

AI is nothing new for print, though – it has been used in various forms, for example in workflow technologies, for many years. More pertinent in print though, perhaps, has been advances in machine learning (ML), which is closely related to AI.

According to Google, ML is a subset of AI “that automatically enables a machine or system to learn and improve from experience”.

This has been used, for example, to predict maintenance that will be required on equipment or breakdowns that are likely to occur, before they happen.

But the media hype over the last few months has centred around generative AI. ChatGPT only launched last November but has already been the subject of much discussion, with its potential uses – and dangers – extensively explored and talked about.

But for now printers are largely unbothered. A recent Printweek poll that asked, ‘Is artificial intelligence (AI) relevant to your business?’ found just 9% of respondents were already using it, while 12% planned to in the future. A further 9% said they didn’t trust AI, while 70% said they preferred the human touch.

It is the latter figure that may put printers off using AI in one of the ways it could be most prominently deployed in print in the near future – for customer service, i.e. to answer phones and converse with customers, or for more sophisticated versions of chatbots that are already widely used on many websites.

Instead of just providing pathways to direct queries to the right department, these systems may eventually be able to carry out full conversations more similar to human interactions.

Ayush Jain, lead software product manager at HP, says the manufacturer has been using various forms of AI and ML in different contexts over the years, including technology that can predict failures of different press components well in advance.

HP uses chatbot technology, he adds, as the first line of contact, as it can currently answer simplistic questions but is not yet at the level where it can replace human interactions.

“Today the chatbot is still a rule-based engine, which is why you’re asked for so many options because it then goes through the different routes. But with generative AI, at least with some of the first experiments that we’ve done, what we’re seeing is that you don’t need a rule-based engine anymore, it is [becoming like] human intuitive conversation and is able to predict based on your previous two or three requests what you’re going to ask next, or what is the depth of information you are comfortable with versus someone else.”

A report by investment bank Goldman Sachs has predicted that AI could replace the equivalent of 300m full-time jobs, and described the ability of generative AI to create content indistinguishable from human-created output as “a major advancement with potentially large macro- economic effects”.

However, the report also suggests new jobs could be created and that AI will increase productivity.

“Some people might look at [AI] as a threat to their position, like it’s targeting writers, PR and marketing people, but at the same time if you flip that, it’s like any other technology,” says Nancy Byrne, VP marketing at Keypoint Intelligence, who suggests creatives could use the technology to help them formulate ideas to save mental energy.

“When it first comes out, those people that adopt it will use it to their advantage, so it becomes a part of their team for more efficiencies and being able to add on additional resources, when you don’t have the budget per se, or to be able to get things done faster. But it’s only going to take it as far as you take it.”

Speaking at the BPIF Print Industry Reception at the House of Lords last month, Unite national officer Louisa Bull also alluded to the threat of AI on industry jobs, and warned that while the sector should not fight change, it needs to make sure workers are “properly equipped”.

Bull told Printweek at the event: “We have to embrace AI, but we need to be sure that everyone is skilled and able to have a job afterwards, or that there’s a society that allows a shorter working week or whatever it is.”

One of the biggest concerns around AI is copyright – who owns AI-produced output and can the output infringe on existing copyrighted material. At the moment, copyright implications are unclear and the subject of much debate, but as with the internet, AI is likely to become more tightly regulated over time.

Oliver Mustoe-Playfair, co-founder of Prinfab, which recently launched Prinfab AI, a free online tool that enables users to create repeating patterns from a text description, says: “Launching an AI product in a creative industry has revealed most people know about AI, but very few people understand how AI works.

“A common misconception is that AI simply copies from the internet, which it doesn’t. AI looks at the internet but learns from it to generate entirely new content. For example, AI can look at a million images of a dog, learn what a dog looks like and generate its own image of a dog.”

Christopher Wheeler, Ricoh head of workflow and RGC solutions, believes that identifying who owns the data and ideation, and whether this can be used in the public domain, is the biggest ‘watch-out’ for printers with AI.

“Copyright infringement is becoming more prevalent and with the continuously changing technological landscape, businesses risk privacy and GDPR implications – particularly when printing in a ‘lights out’ automated environment.

“With the use of AI becoming increasingly prevalent each day across multiple industries, it is vital that the print sector keeps pace with emerging and innovative technologies. As we can see from the Printweek [poll] findings, the results show our industry is unconvinced of AI’s potential use.

“While it may take time to fully impact our sector, we believe print businesses can benefit from AI-generated efficiencies, streamlined processes and shortened production times, but only if they invest time in further understanding AI use cases.”

Wheeler adds that while Ricoh has mainly been piloting the use of AI in the area of customer service in the UK, Ricoh Japan’s development team is actively working to develop AI tools for the wider business, such as IT services, “something that will be particularly useful for data extraction and the reduction of manual touchpoints and documents”.

He suggests that it’s wrong for printers to assume AI is irrelevant to their business and says that many – understandably, given the rate of advancement – don’t yet fully understand its capabilities.

“AI presents an opportunity for the print industry to support critical communications. The rapid and widespread use of AI across multiple industries has harboured fear and scepticism of the cybersecurity risks and the authenticity of online data.”

Another area that could benefit from AI is publishing, and some publishers have already started using generative AI in a limited way to support their work, perhaps for suggestions, prompts or page layouts. Numerous trials of using AI to write articles, however, have found that AI-produced copy is often inaccurate or reads poorly.

Press Gazette reported that Mediahuis is currently producing a ‘Newscondenser’ – an in-house tool that employs generative AI for summarising stories, creating bullet points, and generating headlines for its journalists to choose from.

But Hold The Front Page has recently investigated the story of a local news website in Bournemouth that HTFP discovered was launched with fake journalist profiles and images, and content that the website’s creator admitted was “polished” with AI.

This riled the South West branch of the National Union of Journalists, who warned that such sites could take revenue away from genuine regional and local publishers, and said the case raised questions over the use of AI “in an arena which relies on facts and truths”.

Many observers also feel that, for the sake of transparency and to avoid the spread of ‘fake news’, content generated by AI should be labelled as such.

AI could ultimately contribute up to $15.7tn (£12.3tn) to the global economy in 2030, according to PwC, and its productivity advantages are likely to benefit every industry, including print.

While there are some understandable concerns over the technology, it is worth keeping a close eye on its developments and not dismissing it at this stage, particularly generative AI which is still new and being experimented with.

HP’s Jain concludes: “In five or 10 years from now, generative AI will be an integral part of your operations. You wouldn’t even know that it’s generative AI, but it will be powering a lot of your daily actions.” 


AI is a bucking bronco and we must learn to ride it

Thom Dennis, CEO, Serenity in Leadership 

The writing is on the wall. You may be one of the 79% of Printweek readers who recently said AI is not relevant to their business, but AI is changing just about everything. It is already somewhat out of our control – it is constantly learning and improving, and can write its own prompts and code.

This is a technology in its absolute infancy and yet is spreading and growing faster than anything in human history, and we ignore it at our peril. AI is changing the way print businesses operate, opening up new revenue streams by personalising websites, email, video, and other content to serve customers better. From order processing, print workflow and assessment, the web-to-print storefront is being automated with customers creating customised and print-ready designs. A print run can now automatically notify you when something is out of alignment and then take care of the problem all by itself. AI can also collect and track real-time data to help marketers make smarter and faster decisions.

Intelligent document processing and natural language processing are being used to capture, extract and process data from a variety of document formats without human input. This leads to reduced manual processing, increased accuracy and productivity and improved compliance and security.

The main challenges are potential security and privacy concerns, job losses, misinformation and fake news and a diminution of social connection and empathy, with increased isolation.

The areas most affected by AI will include production workers involved in repetitive tasks, quality control personnel, and administrative roles. On the other hand, AI presents new opportunities for skilled professionals who can develop and manage AI-driven systems, analyse data, and oversee the integration of advanced technologies.

AI is a bucking bronco and we’re going to have to learn to ride it because as far as we can see, we’re not going to be able to tame it.


How do you think AI will impact print?

Les Csonge, business development consultant, BlueToad

“One of the most significant advantages of AI in the print industry is improved accuracy. AI algorithms can analyse large volumes of data to identify errors and correct them quickly and accurately. This can result in high-quality prints, reduced wastage, and an overall increase in efficiency. On a more futuristic note, AI powered virtual assistants could help streamline customer interactions and improve communication between clients and printing companies including dynamic pricing generation, fast research, and the provision of alternative options for clients.”

Simon Biltcliffe, executive chairman, Webmart

“Like every other industry, we’re under huge competitive pressures and, according to McKinsey, 80% of the value created by implementing AI is in four areas: customer service, sales and marketing, research and development, and general innovation. As we’re an innovative customer service driven industry that always needs more sales and marketing, we need to understand and evaluate how we can deliver that into our working practices because otherwise we’ll be left behind. We can’t afford to turn a blind eye to this – it will be more important than the internet to the print industry if we adopt it appropriately.”

Dan Tyler, founder and CEO, Vism

“I think there’s two potential applications for AI in print. Probably over the next few years, the workflow systems that get it right will be able to help printers take advantage of creating a lot more efficiency and putting in place a lot more automation when it comes to the software side of things. On the flipside, we’ve been using some of the more mainstream AI tools in our team project management, coordination, scheduling, and planning. And I think content generation is the use case that we’re going to see more of very quickly.”