How Far Has AI Drug Design Come?

FEATURED RESEARCH

We Tracked 389 AI Drug Candidates to See Which Ones Might Actually Work

Developing a drug takes 10–15 years, costs up to $2.8B, and still fails 90% of the time. AI is changing that. Tools now identify targets, screen molecules, and even design new drugs, in days instead of years. But the line between AI-assisted and AI-developed is starting to blur.

Examples of AI in drug development: BenevolentAI flagged baricitinib for COVID-19 in days, not months. Insilico Medicine’s platform generated fibrosis drug candidates in just 46 days, shaving years off the usual timeline.

AI is also making clinical trials more efficient. It finds eligible patients faster, predicts who’s most likely to respond, and even builds control groups using past patient data.

Considering that nearly a third of late-stage trials fail due to recruitment issues, the hope with AI is that an incremental improvement of success could save years of development time and billions in wasted costs.

But how much AI is enough? There’s still no clear line between AI-assisted and AI-developed. Some drugs are flagged for new uses by AI. Others are optimized or reformulated using AI-driven insights. But most still rely heavily on traditional methods at some stage.

As AI is integrated into more parts of the pipeline, it gets harder to separate where the human ends and the machine begins.

Here’s what we’re doing about it: We’ve built a public database of 389 AI-involved drug candidates from 88 companies (and counting). Not all will make it. But we’re tracking which ones succeed, which ones stall, and how they stack up against traditional methods.

We all know there’s a lot of hype with AI and certainly the progress is very real. But we’re watching closely to see how successful AI in drug discovery really is!

For more details: Full Article AI-Drug Database: Access Here

Brain Booster

Select the right answer! (See explanation below)

Opinion and Perspectives

POST COVID HEALTHCARE

Five Years After COVID, AI Is Reshaping Hospitals and Patient Care

Five years after COVID-19 was declared a pandemic, healthcare looks dramatically different. While the crisis exposed major cracks in the system, it also accelerated innovation, especially in AI and digital care.

Telehealth goes mainstream: Before COVID, telehealth was rarely used. Johns Hopkins had only 80 telemedicine visits per month. By May 2020, they were conducting 90,000 monthly telehealth appointments, and even today, they average around 30,000 visits each month. This rapid shift demonstrated telehealth’s value, particularly in mental healthcare, where virtual appointments improved patient consistency.

AI steps up patient care: Hospitals also embraced AI during COVID, adopting tools to streamline care delivery. Johns Hopkins recently deployed Abridge’s AI technology for clinical note-taking, which clinicians call "transformative," saving doctors over an hour per day.

Doctors are increasingly comfortable with AI; by 2024, two-thirds reported using AI in patient care, a 78% jump from the year before, according to the American Medical Association.

Lingering questions: The rapid rise of AI and telehealth hasn’t come without challenges. The healthcare system still lacks permanent policies for telehealth reimbursement, causing uncertainty for providers and patients. Meanwhile, ethical questions about AI's use, privacy, and reliability remain unanswered.

Moving forward, experts emphasize the need for balanced adoption, embracing the efficiency and access benefits AI provides, while creating clear standards and safeguards for its use.

For more details: Full Article

Top Funded Startups

For more startup funding, read our latest February Report.

Next week, we’ll release the March report!

Product Pipeline

AI IN PATHOLOGY

FDA Recognizes Paige AI for Its Multi-Organ Cancer Detection Capabilities

Paige has received Breakthrough Device designation from the FDA for Paige PanCancer Detect, an AI diagnostic tool designed to assist pathologists in identifying cancer across a wide range of tissues and organ types.

It’s the first AI tool of its kind to be recognized for multi-anatomic cancer detection.

This milestone strengthens Paige’s position as a leader in AI for pathology, helping to address growing diagnostic demands and reduce time to diagnosis.

The tool, available through the Paige Alba™ platform, aims to boost diagnostic confidence, streamline workflows, and get critical results to patients faster.

For more details: Full Article

Policy and Ethics

PEDIATRIC MENTAL HEALTH

Pediatric Bioethicists Say AI Chatbots Are No Substitute for Human Care

AI-powered mental health chatbots are increasingly seen as quick fixes for care shortages, but experts warn that using these tools with children raises serious ethical concerns.

In a Journal of Pediatrics commentary, bioethicists argue that these bots may hinder kids’ social development, bypass family context, and blur responsibility in care decisions.

Most apps are designed for adults and remain unregulated. Researchers stress that children, who form attachments differently and rely on caregivers, need safeguards.

Without inclusive data and transparent oversight, these tools risk reinforcing health inequities.

As co-author Bryanna Moore puts it, “We’re not saying get rid of AI therapy bots. We’re saying we need to be thoughtful in how we use them, especially with children.”

For more details: Full Article

Byte-Sized Break

📢 Three Things AI Did This Week

• By 2030, AI could replace up to 40% of outsourcing jobs in Africa, with women’s roles 10% more at risk than men’s. Experts urge investment in digital skills to avoid worsening gender inequality. [Link]

• Critics say Trump’s new tariffs, some on uninhabited islands, rely on AI chatbot-style math using flawed trade deficit logic. Even the bots warned the approach was oversimplified and risky. [Link]

• New Jersey now criminalizes deceptive AI-generated media, including deepfakes, with up to five years in prison and civil lawsuits allowed—spurred by a student victim who had no legal protection. [Link]

Have a Great Weekend!

Trivia Answer: C) Vampire bat

The vampire bat’s saliva contains desmoteplase, a natural anticoagulant that prevents blood from clotting during feeding. This compound inspired research into new treatments for ischemic stroke, where quick clot-busting is critical.