Global Courant 2023-05-03 08:49:58
But the pandemic clipped its wings; the conference was cancelled.
Ekins and his team continued to do business as usual: finding drugs for rare diseases and those neglected by Big Pharma, who don’t see them as profitable.
His investigations range from “a disease that shortens the lives of children and affects around 100 worldwide” to infections such as malaria because “although in some cases there are medications, over time they cease to be effective.”
“It’s not going to make us rich, but at least I think we’re contributing to society at large,” Ekins said.
A year later, he received another invitation to the conference, although without the attraction of the trip, since it was going to be done by Zoom.
This time Ekins had to pay more attention: it was a meeting to assess new trends in biological and chemical research that might pose security threats.
BBCSean Ekins, chief executive of Collaborations Pharmaceuticals.
“They were interested in the misuse of technology, scientific tools, science in general.”
They were being asked to make a presentation, a request that he found “very strange.”
“It forced me to think about how we could misuse what we do.”
What they do
To discover those medicines against those rare diseases, Collaborations Pharmaceuticals had created MegaSyn, an artificial intelligence (AI) platform.
Although it’s not something you can buy, you can put it together, “just like Lego,” Ekins explained.
“All those little bricks” they needed to build it, they got from “open source software databases”.
“Then we went to databases of information, to create machine learning models that we then connected to MegaSyn; that allows you to design molecules with particular properties.”
That is, they trained her to do the work of a human chemist at inhuman speed.
BBC
Drugs are basically composed of molecules: MegaSyn not only searches among those that already exist, in case one is useful for a disease, but it can create others never seen before.
Dr. Evil
On the eve of the conference, Ekins and Associate Director Fabio Urbina set out to find out what would happen if MegaSyn fell into the wrong hands.
This is how the experiment they called the “Dr. Malvado Project” was born.
It turned out to be simple: when trying to generate a new drug, it is imperative to ensure that it is non-toxic.
What if they tampered with that filter?
“We only moved the direction of a switch on the model.
“Instead of ‘non-toxic,’ we said: ‘toxic.’ It was literally a matter of switching a 1 and a 0 in the program. It’s that simple.”
They hit enter and left MegaSyn to do its thing.
BBCFabio Urbina, associate director of Collaborations Pharmaceuticals.
The next day, they came across a list of tens of thousands of molecules and did a search to see if any were known.
What they found shocked them.
VX
“That very simple change to the program, it had had a dramatic effect.”
With the push of a button, MegaSyn had engineered the terrifying VX nerve agent, which is banned by the UN and classified as a weapon of mass destruction.
It is one of the deadliest chemical substances ever manufactured, less than a drop in contact with the skin is enough to kill the person.
Furthermore, they confirmed that MegaSyn “was able to find previously unknown precursor molecules, starting points towards the development of massively more powerful chemical weapons.”
Getty Images VX molecule.
Most concerning was learning that while Dr. Malvado was a theoretical experiment, others, with other intentions, “could do the same, if they haven’t already.”
“You don’t need a PhD, just some coding knowledge, a basic laptop…everything else is available online for free,” he stressed.
“If some chemist got their hands on this and wanted to weaponize those molecules, since no one knows they exist, those weapons would be impossible to trace and diagnose.”
Dilemma
Alerting the well-intentioned to the danger also meant alerting the malicious to the possibilities.
Ekins and Urbina decided not to give too many details about the outcome of the Dr. Malvado Project in their talk to Spiez.
They wrapped them up in a more general description of their business, giving just enough information to warn the experts.
As camouflaged as it was, it was that potential for negative use of their technology that had the most impact.
Shortly thereafter, an article summarizing the conference talk, published in the peer-reviewed scientific journal Nature Machine Intelligence, caught the attention of newspapers and magazines, including the Washington Post and The Economist, and sparked debate among chemical weapons experts, artificial intelligence, pharmaceutical industry and even philosophers.
Ekins and his team were even summoned to a meeting with the Office of Science and Technology Policy and the National Security Council at the White House.
“It was surreal.”
Getty Images
Despite initial fears that talking about it would serve to inspire rather than prevent, Ekins had to take the risk of raising awareness, not only because of the interest it aroused, but because “he couldn’t sleep in peace anymore.”
But he has never been willing to share that list of tens of thousands of possible toxic molecules produced by MegaSym, not even with the White House.
To do?
Despite everything, no one can stop progress. Platforms like MegaSyn are tremendously beneficial and Ekins knows this better than most.
One strategy to prevent is to “educate people about the problems”.
“Now that the government is aware, we have been invited to speak on the issue on behalf of government agencies.
“In May, for example, we will go to a conference in Brazil, to speak on behalf of the US State Department to academics and industry professionals about the potential issues around generative AI,” he told the BBC. World.
Has this work helped you to sleep more peacefully?
Not much.
Getty Images
“Since we did Project Dr. Evil, generative AI has become much more public, with things like ChatGPT and DALL-E.
“Chat apps have even been shown to be able to design molecules.
“The barrier to entry is much lower today than it was a year and a half ago.
“It is imperative to educate scientists who have not had the same experience as us,” he insisted.
Some experts consulted by the media, however, have insisted that there is no reason to sound so many alarms.
They point out that identifying new molecules is just the first step on a long road toward producing a drug… or a weapon.
And that most of those paths lead to failure.
On both sides of the debate, the same analogy often arises.
The cake
Getty Images
For some of the experts who don’t lose sleep over any of this, what MegaSyn gave was like a list of ingredients.
It would take chefs with all the Michelin stars and a kitchen full of these cooks, they point out, to be able to prepare the perfect cake.
“A year ago, I would have been comfortable with that,” Ekin said.
“There have always been people who have made drugs and weapons illegally, and it’s obviously a niche.
“But we have expanded the potential number of pastry chefs.”
For Ekin, what technologies like MegaSyn produce are not mere ingredients, but recipes that, if they don’t work, help to perfect them.
“Once you provide the recipe, anyone can bake that cake. Some will do better than others, but there is always that danger.”
But there is, he said, something that might help us sleep better.
“We can use those technologies to defend ourselves, use the same tools to find out what chemicals could be used to make new toxic molecules, and then control them.”
Various entities in places like Switzerland are very engaged on the issue, and there is discussion about how bodies like the Organization for the Prohibition of Chemical Weapons (OPCW), which regulates molecules, can fit AI into their legal framework.
Remember that you can receive notifications from BBC Mundo. Download the new version of our app and activate them so you don’t miss out on our best content.
