Abu Dhabi AI-designed painkiller ISM9528 shows preclinical promise vs morphine

What happened

Researchers working with Insilico Medicine, a clinical-stage biotechnology company with scientists in Abu Dhabi's Masdar City, have used artificial intelligence to develop an experimental painkiller that they believe could be more effective than morphine. The candidate, known as ISM9528, is designed to treat acute and chronic pain, and the company describes it as a non-opioid option.

According to reporting by The National, initial results have been promising and the next step is testing in humans. Alex Aliper, president and co-founder of Insilico, said that "to see a non-opioid candidate outperform morphine in preclinical models is unprecedented". Asked about reported side effects, he described the position as "so far, so good".

Insilico presents the drug as a possible alternative to opioid medication, which could reduce the risk of dependence and addiction. Because the comparison comes from preclinical models, the claims have not yet been tested in people.

Why it matters

The story is less about a single molecule than about what AI now makes possible in discovery pipelines. When a model-driven process produces a candidate that its developers regard as ahead of an established comparator at preclinical stage, it changes the pace and economics of how organisations design and test products. For leaders in AI and digital transformation, the signal is that AI-driven tools are moving from support roles into core research workflows. That shift requires the operating model, data practices and validation steps to be redesigned around them.

The development also shows AI being applied in a regulated, high-stakes setting. Developments of this kind are likely to raise questions about how new tools are governed, validated and trusted by the people who depend on them.

By the numbers

• Nine in 10 potential drugs that reach the human testing stage fail to be approved, according to health experts cited by The National.
• ISM9528 is the candidate in question, with human testing as the next stage reported by The National.

The Renascence take

Most readers will focus on the morphine comparison, but the more useful signal is the process behind it. The headline claim is preclinical, and the sector's own attrition rate at human testing is high, so this is a reminder that AI can accelerate the front end of a pipeline without removing the scrutiny needed at the back end.

The behavioural principle here is the gap between a promising signal and a trusted outcome, and it applies to any customer-facing AI launch as much as it does to a drug. Operators should treat early results as a hypothesis to test in real conditions, not a verdict. The practical move is to define success criteria and failure thresholds before a pilot begins, and to be candid with users about what has and has not yet been proven. Teams that do this keep trust intact even when the first version falls short.