An eye on AI

Since ChatGPT was launched back in November 2022, the topic of Generative AI has become an almost permanent fixture within the media.

From exploring its transformational impact on virtually every industry, to the potentially negative consequences of the technology, the debate is – seemingly, never-ending.

While a healthy level of scepticism is always encouraged, it would be a great shame if AI’s reputation were to be damaged to the point that the public lost trust in the technology altogether.

Of course, the pharmaceutical industry must tread carefully.

The products being developed and manufactured are responsible for saving lives, and no corners can be cut during the production, labelling or distribution process, but shunning AI altogether certainly isn’t the answer.

With the right processes in place AI can do amazing good, and we shouldn’t forget that.

Animal testing was, sadly, an uncomfortable reality for the cosmetics industry for many years, until companies had enough data to reliably understand each ingredient and confidently say that their products were safe for human consumption.

You can draw parallels between this example and the use of AI. Of course, there are risks to consider and if those seeking to do harm got hold of the technology, they could cause damage.

However, AI can also provide businesses with a wealth of insight, speed up processes and help them navigate regulatory changes more effectively.

That final point is essential, given that the Windsor Framework, a new set of regulations designed to protect the Good Friday Agreement, is coming into force at the start of 2025.

These new regulations will require thousands of product labels to be updated on goods moving between Great Britain and Northern Ireland.

As you can imagine, that’s no small task. Many of these products are used across the NHS on a daily basis and if these labelling and artwork updates can’t be processed quickly enough, there will undoubtedly be delays and temporary medicine shortages.

There is a solution that uses AI to analyse vast data sets to identify key terms within the metadata of each label – saving hundreds of hours of time when labels need updating in line with new regulations.

Those attempting to manage this process manually will have to source information from the marketing and sales teams about where each product is sold and distributed.  Often these departments work in silos, making the challenge even greater.

They would then have to manually search potentially tens of thousands of labels to identify which ones need updating. It’s easy to see how this process could lead to shortages of certain products simply due to the amount of time it would take.

Another benefit of AI is that it learns as it goes. If the AI makes a mistake, you can tell it to not follow that logic again to prevent the same error being repeated - hence, the term ‘Generative’ AI.

Of course, when a business first starts using AI for a certain application more input is required, however this level of input reduces over time.

So, despite the growing stigma around AI, these examples demonstrate how effective it can be in getting treatments to patients much more quickly and, ultimately, saving lives.

As well as navigating regulatory changes, it’s also not unheard of for certain pharmaceutical products to suddenly be in huge demand. Ozempic, a drug manufactured by Novo Nordisk, and initially used to treat diabetes in 2017, is a prime example.

Having been discovered as an effective weight loss drug, used by celebrities to get ‘red carpet ready’ and subsequently going viral on TikTok, demand quickly surpassed even the manufacturers’ expectations, leading to severe supply issues.

To capitalise on this, the company is building an additional manufacturing plant, but even with this added capacity, there’s still the labelling and artwork to consider.

Any new artwork will need to be submitted to the regulator for approval and, without an automated solution, this process would take weeks and could lead to thousands in lost revenue.

We saw a similar situation when we worked with Novavax in March 2022, on its COVID-19 vaccine labels. At the time, it told us that getting the labelling and artwork ready and approved would ordinarily take it between eight and ten weeks.

Given the critical need to get this drug out to patients, Kallik’s technology was deployed – helping cut this process down to just four or five days.

There’s clear evidence that AI can be used as a force for good when deployed responsibly.

It’s unlikely that any political changes in the UK will impact regulations as medicines continue to be regulated by Europe.

Instead, the priority over the next few years should be around building trust that AI can – and will, help to save lives.

Bob Tilling is VP Global Sales at Kallik.
Go to kallik.com