The Critical Need for AI Literacy in Pharma Leadership

We’ve said it before, in our article Hybrid Intelligence: How AI roles can lead to commercial success (or failure), “What’s important to understand is that AI project failure doesn’t occur in a vacuum. People are more than just secondary considerations to hardware and software when it comes to AI — the reason things don’t work out as they should is almost always down to human rather than technical failure.”

Why is AI literacy so critical for pharma leadership?

Making people primary considerations in pharma’s use of artificial intelligence (AI) involves ensuring AI literacy in leadership and across boards, not just in our technical managers and data scientists. The lack of AI literacy in leadership sits behind the reasons why most AI initiatives fail. AI literacy in leaders– especially in non-technical leaders – also powers AI success. We need AI literacy in the C-suite to develop strategic initiatives that have the potential to succeed to move the needle on real world impact from AI initiatives.

The staggering statistics: AI initiatives usually fail

It’s nothing new, and we’re already too immune to the fact, but most AI initiatives fail.

As Gartner revealed, as far back as 2018, “85% of AI projects will “deliver erroneous outcomes.” Other surveys and reports come out with similar statistics with the failure rate of AI projects between 83% and 92%. The fact that only two out of five businesses report gains from their AI investments reflects this. A mere 13% of data science projects making it into production provides evidence of this too. The global AI pharma market expects to reach $3,624.24 million in 2026. That’s a shocking waste of money if we continue with such high failure rates.

C suites know they want and need artificial intelligence in their pharma businesses – so they are spending so heavily. They know it has the potential to transform their future success and deliver unrivalled competitive advantage. So they throw money at AI, believing that simply doing so will lead to pharma success. But when seven out of 10 executives who have invested in AI report they’ve seen minimal or no impact, and failure rates are so painfully high, what’s going wrong?

Why do AI projects fail?

There are multiple reasons cited for the lack of consistent AI project success. Each of these reasons holds merit. However, one overriding factor behind each of these reasons is low AI literacy at the pharma board and executive leadership level. In contrast, the need for AI literacy amongst this population is high. Pharma leaders need to understand AI far, far better than they currently do.

Let’s look at some common reasons why artificial intelligence projects in pharma tend to fail and how those reasons wouldn’t be present if there were greater AI literacy at the leadership level.

·      AI initiatives are oversold

It’s easy to oversell AI solutions because of the temptation to think of AI as magic. We can blame vendors of AI solutions for pharma for stating that their platform does more than it does when in reality, the pharma leaders didn’t understand its capabilities and expected more than it could deliver.

How AI literacy in leadership would eliminate this: A clued-up C-suite won’t fall foul of seeing AI solutions as magic if they understand enough about how and why AI solutions work. They won’t fall for the sales spin but will have realistic expectations.

·      Data isn’t understood as the bedrock

Any AI solution is only as good as the data it sits on. Even the most technically capable AI solution appropriately utilised won’t deliver if the data behind it is poor. The big whammy of AI project success is clean, accurate, prolific, and mineable data.

How AI literacy in leadership would eliminate this: Right now, even understanding that the data is everything is missing from leadership understanding. AI literacy gives data the importance it needs in any AI project. Beyond this, leadership should have the ability and foresight to conduct a data audit and feasibility study first.

·      The solution is seen as more important than the strategy

The belief that AI is magic is tied closely to the idea that it can solve all problems. Often AI projects are approached with shiny things disease rather than identifying the problem first. It’s why an AI blueprint is so important; the solution should first solve a defined problem identified from wider business conundrums. Without this, AI developers and vendors create solutions that fix their perceived understanding of your problem rather than meet the objective and desired outcome.

How AI literacy in leadership would eliminate this: AI literate leaders understand the importance of defining the problem that needs solving so that objectives are clearly understood. They can communicate expectations clearly to the AI development team because they know enough about the capability of AI, but importantly, they understand enough about the presenting corporate challenges that need to be solved.

·      There’s a misunderstanding about who does what

Developing AI solutions is like a newborn baby regarding organisational structure because we don’t have a set format for who does what on AI projects, and we’re still learning. At best, we approach it as another IT project, which is woefully short of what it is. As such, projects flounder across silos. No one talks to anyone else because they don’t even realise they should.

How AI literacy in leadership would eliminate this: An AI literate C suite would stop the silo effect in artificial intelligence projects because it would see the need for a dedicated and collaborative team that pulls together the full spectrum of skills and know-how in a perfect blend. Like we know the different components of an effective clinical research team because we’ve been using them for decades, AI literacy would allow boards to understand how to create effective AI project teams.

·      Underestimating what needs to happen for AI buy-in

It’s one thing to develop an AI solution, but it’s another to get your people to use it. It may be all-singing-all-dancing, but it’ll fail if your people aren’t on board culturally. Partly this is a matter of training and talent acquisition, and partly it’s a matter of cultural shift.

How AI literacy in leadership would eliminate this: Leaders well versed in AI literacy know that this is a problem and head it off by effecting cultural shifts from the top and underpinning the importance of understanding for all. Leaders must remove the black box effect from AI implementation, whereby users don’t understand why the solution works and become suspicious of it.

What does AI literacy look like for leadership?

No one is expecting pharma leaders to become experts in developing AI solutions. We don’t expect C suites and executive leadership in pharma to know their Generative Adversarial Networks (GANs) from their dimensionality reductions. They don’t need to know anything about Python, Lisp, Java, Julia or even Haskell.

However, pharma leaders will benefit from becoming AI literate and becoming experts in how to get what they need from AI vendors, data scientists and the like. This means knowing about AI capability – they need to understand how and why AI works. AI literacy involves knowing why an AI strategic blueprint is essential and how to create one. Furthermore, with AI literacy, leaders must understand why and how AI is already succeeding and how to mirror that within their business for their own specific strategic challenges and objectives. It involves understanding the data available and the problems that need fixing. Finally, it includes understanding how AI fits in with broader business goals.