Does the potential of AI to accelerate drug discovery really require a reality check?

As I write this post, I'm also building Future Therapeutics , where we're developing and leveraging proprietary AI infrastructure for end-to-end therapeutics discovery. 

The underlying idea is to eliminate silos and streamline the entire process.

And when all is said and done, the aim remains constant: to bring more new cures and treatments for life-threatening diseases in less time and while spending fewer €€€.

The emerging question that however remains unanswered is: Does the potential of AI to accelerate drug discovery require a reality check?

And I kind of agree with it that it's the time.

But let me clarify; 

While I’m not an expert (yet) to talk about the intricacies of clinical trials, there are plenty of other experts who can fill in those details. 

I can attest to AI potential in the discovery and preclinical stages—it undeniably accelerates processes and saves both time and money!

I've experienced it firsthand during my PhD (2016-2021) at The University of Freiburg and continue to witness the remarkable superpower it brings.

And if you're someone who appreciates numbers, consider this: according to BCG, AI can lead to "time and cost savings of at least 25–50%" in drug discovery up to the preclinical stage. 

Now, that's impressive efficiency! Isn’t it?

And, for those who value numbers in $$$, Morgan Stanley Researchers have simplified it even further.

They suggest that even modest improvements in early-stage drug development success rates, with the help of AI, could yield an extra 50 novel therapies over a decade. 

This could potentially represent an opportunity surpassing $50 billion.

What's equally astonishing is that despite AI technology being relatively new, the 2022 BCG analysis revealed that 20 AI-intensive companies collectively presented 158 drug candidates for discovery and preclinical development between 2010 and 2020. 

Yes, you read that right—158! 

And with ongoing advancements, I can only imagine that this number has soared even higher since then.

And get this: the world’s 20 largest pharmaceutical giants, measured by revenue, had ONLY 333 candidates in the same period.

So there's no doubt that AI holds immense potential to accelerate drug discovery. If done correctly, it can bring revolutionary changes in advancing new cures and treatments to patients in need.

So why are both I and the October 2023 Nature Editorial still suggesting that it needs a reality check?

Transparency is the biggest cause for it.

Ultimately, these claims originate from the companies themselves.

When was the last time you heard or saw a company openly discussing their challenges?

Discussing the technology struggles with generalizing from one target/disease class to another?

Resorting to conventional approaches in the project pipeline, yet openly acknowledging it?

In the end, everyone is revolutionizing drug discovery with AI :)

However, when everyone is doing it without being transparent, it becomes very difficult to discern what is actually working and what is not.

Being an ardent believer that leveraging technology is key to discovering and bringing life-saving treatments to market on time. In the current technology landscape, AI is our best option! 

I couldn't track down the author, but I echo the sentiment expressed in the October 2023 Nature Editorial that these claims should be independently verified. 

Transparency can have a profound impact and bring all stakeholders closer together, enabling us to recognize challenges, identify gaps, and work collaboratively toward finding cures and treatments for the countless patients in need.

Let me know what you think.