Manufacturing with Al

Robot's head in face
Robot's head in face
Robot’s head in face

OpenAl’s latest technology for creating artificial realities has caused a significant buzz in the market with the announcement that Sora can produce realistic and imaginative scenes from text instructions.

This could be a step change for the manufacturing industry, accelerating productivity, reducing expenses and downtime, and improving product quality.

Large manufacturers aren’t the only ones who stand to benefit from this technology – smaller businesses struggling to get access to high-value Al solutions can ride on the success of Sora’s bold and evolutionary capabilities to enhance performance and production schedules.

A manufacturer could take a drawing off a page and show it in motion, in real-life situations and operate it to perfection long before any physical product is built. If this allows the manufacturer to test the market and generate demand, it has the potential to democratise time-to-market and customer acquisition.

In some ways, Sora is an incremental step – given that image generation has existed for a while with Google’s Lumiere and Meta’s Make-A-Video. But Sora brings a whole new dimension with its photorealistic videos. It represents a significant reset in terms of information immediacy and value and is certainly hastening advancements in Al-generated videos.

As video capability continues to evolve, immersive videos are also within reach, opening more possibilities with augmented reality, and for manufacturers to take advantage of the metaverse. Using generative Al technology, manufacturers can run virtual models of physical facilities, incorporating augmented reality, the Internet of Things and cloud computing.

Plugging gaps in manufacturing

The real positive of Sora is in responding to some of the biggest manufacturing challenges, such as worker and skills gap and the pursuit of quality excellence. In Australia alone, there is a growing skilled labour shortage in the manufacturing industry, driven by technology shifts, an ageing workforce, and global competition.

Addressing the worker challenge is critical to build up the local manufacturing ecosystem – task-specific training content is a fundamental part of the solution. There is a move to develop training modules to address the lack of talent and upskill professionals with the right tools to successfully scale up operations.

Unfortunately, the manufacturing sector lacks worker capacity and skills to produce the content in a timely and cost-effective manner. This is where Sora steps in to use text to create realistic and imaginative training videos to support upskilling and quickly respond to lost experience. It can generate complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background.

According to a report by Microsoft and the Tech Council of Australia, generative Al could contribute as much as $5bn annually to Australian manufacturing by 2030. Furthermore, McKinsey and Company predicts that generative Al has the potential to increase Australian labour productivity by 0.1 to 1.1 percentage points a year through 2030. This demonstrates a real opportunity for manufacturers to harness this tech to bolster overall business competitiveness.

Regulatory and ethical concerns

In an age of misinformation, malware and fake news, there are growing concerns around how Sora could potentially fuel video deepfakes. There are also fears that if the video creation process allows a product to perform beyond expected capabilities, presumably but not always with a disclaimer, then the customer could be easily and deliberately misled.

A second part to the potential hazard is the creation of defamatory content – a competitor, their agent, or even a dissatisfied consumer having a quicker method to create negative content about a product or individuals within an organisation.

Despite the potential drawbacks, there is no denying that Sora offers a faster and less costly way than the current human-led approach using computer-generated imagery, potentially encouraging the worst of behaviours.

The only way to overcome any misuse is the same as for generated images – providers of public services must prohibit the use of copyright and private input and, in all cases, watermark the output so that the ability to trace back to the originator acts as a deterrent.

Significant opportunities for Australian manufacturers

There are obvious challenges with rapidly evolving technologies, and there is a need to balance innovation with responsibility. Legislative bodies will need to keep up, including ensuring that the industry adopts ethical practices. Once the industry accepts and moves on from those concerns, the benefits of Sora to manufacturers in democratising media and mitigating training and quality challenges of today will be fully realised.

Sora represents the opportunity to educate and excite, and ultimately is a worthwhile prize for those organisations that have adopted the transformational foundations to embrace this future. For Australian manufacturing, it marks a new age of productivity and innovation – with the potential to enhance business competitiveness and boost economic growth.

By Matthew Addley, Industry Strategist, Manufacturing at Infor

This article was first published by AMT magazine