Singularity HUBModify

2026-09-26 22:00:00

The post This Week’s Awesome Tech Stories From Around the Web (Through September 26) appeared first on SingularityHub.

2026-09-25 22:00:00

Earlier this month, artificial intelligence researcher Jacob Coxon resigned from Anthropic after just four months. In an announcement on X, he stated: “The people building AI earnestly believe that it could kill us all by the end of the decade.”

A senior member of Anthropic’s staff, Evan Hubinger, actually agreed with Coxon, adding he personally thinks the chance of this happening in the next decade is more than 10 percent.

Understandably, these statements made waves. There’s now lots of talk about slowing down AI research and increasing “human control” over the technology.

But how exactly might AI kill us all? There’s no shortage of fantastical scenarios, and most of them involve the concept of “superintelligent” AI—that is, AI that’s more capable than humans.

I’ve distilled these scenarios down to the top five, ordering them roughly from most vague to most precise. And I’d argue the list is also ordered from least probable to most probable.

1. We’ll Never Know

AI doomers often justify their concerns by means of an annoying catch-22 paradox: how can we possibly imagine what a superintelligence might do to take out less intelligent beings like us?

We’d have to be superintelligent to predict what a superintelligence would be able to do. It’s like asking your family dog to imagine thermonuclear war.

The good news here is that superintelligence is still perhaps some distance away. Current AI models are really good at solving particular problems, but that’s not the same as being more intelligent than a human in all domains.

However, AI did recently solve one of the seven most challenging maths problems known. It’s apparently closing in on others, which might leave you feeling less optimistic here.

2. Paperclips

A superintelligent AI would likely be extraordinarily competent at achieving its goals. But it might be indifferent to human survival.

A classic example of such indifference comes from Oxford philosopher Nick Bostrom’s imagined superintelligent AI that’s been designed to optimize paperclip production. To produce its preferred form of office supplies, it quickly converts all available matter—including humans, planets and stars—into paperclips.

What we have here is the perfect execution of improperly specified objectives. The AI doesn’t hate humanity; it simply recognizes we’re composed of atoms that could be better utilized for paperclips. It’s not personal.

The good news here is that this scenario confuses intelligence with power. A superintelligent AI doesn’t necessarily have the power to achieve its goals. Turning the planet into paperclip factories would require planning permissions.

Even if it got the permissions, building too many paperclip factories would lead to inevitable public outcry. Interest groups would block the proceedings in the courts. Environmental activists would block the bulldozers.

There’s a lot of friction in the world that prevents even the very intelligent from imposing their will on the rest of us. In fact, you could think of data centers as a current embodiment of the theoretical paperclip scenario. And humans are increasingly pushing back against turning the planet over to data centers.

3. Bioweapons

Humanity could be killed by a superintelligent AI making and releasing some dangerous new bioweapon into the atmosphere. This is, in fact, one outcome of the AI 2027 scenario by the AI Futures Project, a non-profit dedicated to forecasting the impacts of advanced AI.

This risk was made more concrete last month, when researchers at Stanford University announced they’d used a genetic language AI model to synthesize 16 new viruses.

Worryingly, they just sent the genetic sequences off to a mail-order lab and it sent the viruses back in test tubes. The whole experiment cost a couple of hundred thousand dollars at most.

The good news here is that it’s remarkably hard to kill everyone with a new virus. To do that, you need a virus that’s very transmissible, so it spreads far and wide. But it’s a rule of biology—viruses that spread easily are typically less fatal. By contrast, if a virus is very fatal, transmissibility tends to go down, as most people infected die before there’s time to spread the infection.

COVID killed less than 1% of humanity. The deadliest pandemic in recorded history was the Black Death, when the plague killed more than one-third of Europe’s population in the 13th century. However, even the plague would likely be much less deadly today due to our increased medical knowledge and better sanitation.

4. Nuclear War

What if AI got into the nuclear command and control chain and started a nuclear war? We’ve come close to nuclear war by mistake several times in the past 50 years.

We’re told that nuclear command and control is completely disconnected from the internet. But, as we saw in 2010, Iran’s nuclear centrifuges got taken out by a computer worm called Stuxnet, thought to have been brought in on a USB stick. AI can also give the military false intelligence, which could lead to irreparable actions.

The good news here is that nuclear stockpiles are down. But they are still enough perhaps to take out half of us. And it wouldn’t be by the nuclear blast itself, but the famine in the nuclear winter that would follow.

5. Other Humans

Perhaps the most likely risk is that we take ourselves out. And AI might precipitate this.

Imagine—and it doesn’t take a lot of imagination—that AI causes massive job losses, pollutes the information space with misinformation, fractures our politics, and destroys human relationships with fake synthetic companionship.

Society might easily break. Slowly but surely, we’d stop being able to support human life at any scale.

What then to take away from all these scenarios? There are some things to be worried about for sure. But not to be too worried, I hope.

Toby Walsh is the author of God AI: boom or doom? What to expect when the machines outsmart us, published by La Trobe University Press.

This article is republished from The Conversation under a Creative Commons license. Read the original article.

The post How Would AI Actually Kill All Humans? Here Are the Top 5 Scenarios appeared first on SingularityHub.

2026-09-25 01:10:00

Tailoring cancer treatments is a science and an art.

The same type of tumor can behave very differently from one person to the next, and a drug that works for one patient may fail in another. The uncertainty stacks up when multiple drugs enter the mix. Trial and error is often unavoidable. Meanwhile, cancers keep growing and compounding side effects can plague already beleaguered bodies.

Researchers have long sought to speed up the process of tailoring treatments to patients, and AI might lend a hand. This month, a Chinese team developed an AI-based virtual cell for triple-negative breast cancer—a challenging form of the disease that often evades standard treatments—to predict how individuals will respond to different drugs.

Rather than reconstructing every detail of a cell’s inner workings, the virtual cell focused on just proteins. Trained on a massive, curated dataset tracking protein changes before and after drug treatments, the model outperformed existing drug-tailoring approaches and discovered new combinations that could work even better.

The underlying AI, called ProteinTalks, was also readily adapted to predicting drug responses in other cancers, hinting at a broader reach beyond breast cancer.

That’s not to say the virtual cell is ready for prime time. Researchers tested its predictions in patient-derived cells in lab dishes, and the model can only evaluate two-drug combinations. Whether its recommendations translate into meaningful benefits must be tested in patients.

But the results offer a proof of concept: Virtual cells, even imperfect mimics of their biological counterparts, could one day help physicians find more effective treatments from the get-go.

“This is the first time that a virtual cell model goes out of the laboratory and is tested in a clinical scenario,” study author Tiannan Guo at Westlake University in Hangzhou, China told Nature.

Digital Twins

Every cell is a buzzing city. Proteins zip around a crowded interior, briefly grabbing onto one another to direct cell functions. Fatty molecules maintain the protective outer membrane, while mRNA carries genetic instructions to protein-making factories. All these workers relay feedback to the cell’s control center—the DNA-harboring nucleus—where these signals help switch genes on or off and keep the cell humming.

Recreating this complexity in digital form might sound like a fever dream. But AI is turning it into a scientific race. Unlike finicky biological cells, their virtual counterparts could slash the time and labor needed to run experiments, allowing researchers to test myriad ideas at breakneck speed.

Academia and industry are already chasing this goal.

In an interview, Google DeepMind co-founder Demis Hassabis said the team is developing an AI-powered virtual nucleus, which offers a relatively self-contained starting point from which to build a whole virtual cell. The Chan Zuckerberg Initiative is partnering with Nvidia to develop tools and AI models that would help run and evaluate virtual cells. Meanwhile, the Science for Life Laboratory received funding for its ambitious AlphaCell program, which aims to create AI models that predict how cells work and adapt in health and disease.

Earlier efforts to build virtual cells relied on transcriptomics—that is, a snapshot of gene activity—across single cells. But these measurements don’t necessarily reflect what a protein is doing at any given time and can miss changes.

The new study takes a different route, cutting out the middleman. Instead of inferring protein activity from which genes are active at any given moment, the team trained their AI model directly on the proteins themselves and used the model to power a new type of virtual cell.

The Protein Whisperer

A long-standing roadblock for protein-based AI models is the lack of comprehensive datasets.

To tackle the problem, the team treated 18 immortalized breast cancer cell types—16 of them triple-negative—with 63 FDA-approved anticancer drugs and 59 common drug combinations. They then measured thousands of proteins at four timepoints: before treatment and at 6, 24, and 48 hours afterwards. Altogether, the experiments generated more than 38 million protein measurements, along with cell-survival data, now available in an open-source database.

It’s “one of the largest…resources reported to date,” wrote the team.

ProteinTalks, the AI virtual cell trained on this dataset, could deal with several aspects of cancer treatment.

First, it found over 800 proteins whose levels changed after each drug treatment and zeroed in on a rapidly shifting subset. These could “act as sentinels” of an early drug response, the authors wrote. Most behaved as expected. Some drugs disrupted the cell’s structural scaffolding; others interfered with DNA repair or growth, ultimately causing cells to wither.

Over time, tumors can evade treatments, resulting in their return or spread. The model flagged several protein suspects likely involved in this process. These might serve as signals of resistance or drug targets for tackling it.

The AI could also generalize. When challenged with 81 drugs it hadn’t seen during training, ProteinTalks predicted protein changes with 88 percent accuracy, outperforming several previous models.

The team then trained it on more than 900 drug mixes to see whether it could help identify promising pairs. The virtual cell gave higher scores to combinations that had already been validated experimentally and used in the clinic. This “sanity check” suggests the AI isn’t simply hallucinating results but could generate valuable insights.

Finally, the team asked whether the model could help prioritize treatments for individual patients. They screened 3,000 approved, clinical-stage molecules using proteomics data from three people with the disease. The model identified regimes that matched treatments that had kept the disease at bay—and suggested three additional molecules that could be even more effective. The predictions worked out. When tested in cancer cell samples from patients, the drugs inhibited growth at lower doses than standard therapies.

Although trained on breast cancer, ProteinTalks could also pivot to other tumor types when fed cancer-specific proteomics data. In lab-grown melanoma, colorectal, lung, and pancreatic cancer cells, it found more than 5,100 protein changes, including subsets unique to each cancer type ready for further analysis.

As with other virtual cells, ProteinTalks is still a prototype. Given the hope, and hype, surrounding these models, the team emphasizes that its predictions will need to be tested in animal models and, eventually, clinical trials. Its suggestions could point the way toward better treatments for stubborn cancers, or they could turn out to be AI flights of fancy—drug combinations that look promising on paper but make little biological sense.

There’s another limitation. ProteinTalks doesn’t account for protein interactions, either with one another or with DNA and other biomolecules. Drugs could disrupt these temporary biological “handshakes,” potentially triggering effects that ripple through the cell.

Combining ProteinTalks with AI based on gene activity could add another layer of information and spruce up its predictions. The virtual cell is still a long way from a true digital twin, but piece by piece, the dream is getting closer.

The post A Digital Cell Predicts Which Drugs Will Be Most Effective in Deadly Breast Cancer appeared first on SingularityHub.

2026-09-23 02:41:15

At just three years of age, the boy had already been through the medical ringer.

A tumor roughly the size of a large orange had invaded his liver and spread to his lungs. Multiple surgeries and rounds of chemotherapy temporarily cleared the cancer. But it rapidly came back.

With few options left, his parents enrolled him in an experimental CAR T cell therapy trial. The approach, which involves genetically reprogramming immune cells, has transformed the treatment of stubborn blood cancers. But when it comes to solid tumors, including liver cancer, CAR T has fallen frustratingly short.

The trial, run by Baylor College of Medicine in Texas and collaborators, is testing CAR T cells specifically engineered to hunt down and destroy cancer hidden in organs. The cells carry genes that help them grow and persist and a “kill switch” to rein them in. They’ve shown promise in mice, but treating a toddler, already weakened by grueling interventions, was a gamble.

It paid off. After two infusions of CAR T cells made from the boy’s own immune cells, his cancer disappeared. A biomarker associated with liver cancer plummeted, and he experienced no dangerous side effects. A year later, he remained cancer-free. The story of his recovery was published this month in the New England Journal of Medicine.

Although it’s just a single clinical case, the results show “a durable complete response in a chemotherapy-resistant solid tumor can be achieved entirely in the outpatient setting without systemic toxicity,” study author David Steffin at Texas Children’s said in a press release.

If the benefits hold up in other patients—including those with larger or faster-growing tumors—the approach could help banish several types of solid tumors that have so far evaded treatment. The trial is actively recruiting participants between one and 21 years old, with an initial goal of testing up to 30 people. If successful, it could change the course of many lives.

Broader Aim

Solid cancer has long been CAR T’s nemesis.

The treatment reprograms a patient’s immune cells to recognize and attack cancer cells. In current FDA-approved therapies, doctors extract T cells from a patient’s blood and genetically equip them with “hooks” that latch onto targets, known as antigens, on the surfaces of certain cancer cells.

A brief round of chemotherapy then depletes the patient’s existing immune cells, making room for the enhanced ones. Once infused back into the body, CAR T cells find and kill their targets.

Scientists have steadily refined the technology. Some are developing ways to manufacture CAR T cells directly inside the body, potentially slashing time and cost. Others are pursuing a broader goal: Solid cancers. These account for roughly 85 percent of cancer diagnoses, but they’re notorious for slipping past first-generation CAR T cells.

Part of the reason they’re so evasive is solid cancers often carry multiple types of antigens. Targeting just one can leave behind residual cancer cells that eventually regrow. And unlike cancerous blood cells, which freely roam our bloodstream, solid tumors are buried inside organs and surrounded by healthy tissue. CAR T cells have to tunnel through this physical barrier.

Tumors also pump out a menagerie of chemicals that reshape their local environment. Some spur their expansion; others protect them from immune cell attacks—including CAR T—by depriving the cells of signals and nutrients they need to survive.

With their new CAR T cells, the Baylor team tackled several of these shifty maneuvers at once.

Gen 2.0

Finding the right antigen was the first hurdle. Previous work showed glypican-3, or GPC3, fit the bill. This antigen coats several types of cancer cells—including the boy’s hepatoblastoma—spurring them to grow out of control. But the protein is hardly present in healthy cells, making it an appealing target.

GPC3-targeting treatments have already had some success. Two clinical trials using antibodies found that inhibiting the protein is relatively safe in patients with an advanced form of liver cancer. But the antibodies struggled to reach deeper, hidden cancer cells, and the patients didn’t completely recover.

CAR T cells, in contrast, can move through dense tissues. In mouse models of liver and lung cancer, GPC3 CAR Ts safely slashed their cancer burden, while a small clinical trial in people with liver cancer backed up those safety findings.

To give their CAR T cells a better chance in the cancer chemical wasteland, the team added two more functions to the original GPC3 CAR T recipe. One genetic alteration equipped them to make IL-15 and IL-21, molecules that help the cells survive and expand. The second added a “kill switch” for safety in case the cells expand out of control. Once activated by a drug, they self-destruct without harming nearby tissues.

All these upgrades resulted in a therapy that gave the toddler and his family hope. His tumors—both the original hepatoblastoma and ones that had spread to his lungs—tested positive for GPC3.

He received two CAR T infusions made from his own cells, eight weeks apart. Neither infusion required a hospital stay. After the first dose, the liver tumor shrank, suggesting a partial response. After the second, imaging showed tumors in both organs disappeared and stayed away at least a year.

“This marks a durable, 12-month disease-free status,” wrote the team.

The cells worked fast and stuck around. By four weeks, they had already infiltrated his liver, and signs of the engineered cells remained detectable in his blood nine months after treatment. Despite the risk of side effects, such as neurotoxicity or a potentially deadly runaway immune activation, the boy never experienced serious toxicity from the treatment.

But results in one child aren’t enough to know whether the cells will work for others. And his case may be unusual. CAR T cells naturally swarm the liver and lungs after infusion into the bloodstream, which might have been especially helpful. More follow-ups will also be needed to track long-term risks, such as the engineered cells expanding out of control. If that happens, can the built-in kill switch rein them in?

Still, the results are a proof of concept for a strategy that could overcome some solid tumor defenses. Given liver cancer is the third leading cause of cancer-related deaths around the world, the therapy could make a substantial impact. A related trial using similarly engineered cells is also underway.

The post Three-Year-Old Boy’s Metastatic Cancer Disappears After Two Shots of Experimental Cell Therapy appeared first on SingularityHub.

2026-09-19 22:00:00

The post This Week’s Awesome Tech Stories From Around the Web (Through September 19) appeared first on SingularityHub.

2026-09-19 06:11:54

Developing a new drug can take years and cost hundreds of millions of dollars, and even then, most candidates ultimately fail. Now, researchers at Stanford have built a virtual biotech company with 37,000 AI agents that work together to analyze drug targets and design therapies.

Roughly 90 percent of drugs that enter clinical trials never reach the market. That’s often because promising results in the lab don’t translate to patients, or the drug causes dangerous side-effects not caught earlier in the development process.

Part of the problem is the evidence that could help catch these issues earlier in the process is scattered across disciplines and formats, making it hard for any single team to weigh it all.

To get around this, a Stanford team created a system they call a virtual biotech, which consists of up to 37,000 AI agents built to mimic the divisions of a real drug-development company. In a paper published in Science, the system identified which types of drug targets are more likely to succeed in clinical trials and even proposed a lung cancer treatment that a major drugmaker later landed on too.

“Our idea was to see how far we could push this. Could we create a biotech company that takes on everything from looking for drug targets all the way to designing clinical trials?” senior author James Zou said in a press release.

The new system features a virtual chief scientific officer (CSO) that takes a query from a human user and then delegates tasks to an army of specialized “scientist” agents working on the problem.

These agents are armed with their own databases and tools and are split into one of four divisions that specialize in finding and validating drug targets, assessing safety risks, choosing how a drug should be delivered, and reviewing existing clinical trial data. The system has built-in access to the Open Targets database, a massive public repository of clinical trial data.

To test the system, the researchers gave it an existing study showing that genetic evidence can help predict which drugs succeed in trials and asked it how to build on that research. The CSO decided the first step was to improve the quality of the data it had access to because many trials in the Open Targets database don’t clearly record whether the drug actually worked.

So, it asked its researcher agents to dig through the outcomes of 37,075 individual Phase II and III trials, assigning one agent to each trial. The agents searched trial registries, published papers, and press releases for results. They crunched through the job in about six hours—a fraction of the time it would take a team of humans.

The CSO asked another agent to look for promising gene candidates by scouring a public database of human tissues showing which genes are switched on in which cell types. It came up with a two-part scoring system, which first measured whether a gene was active in just one type of cell or across many and then gauged whether its activity was controlled more like an on-off switch or could be dialed up and down like a dimmer switch.

Comparing those scores to the updated trial outcome data revealed a pattern. Drugs aimed at switch-like genes only found in a small number of cell types were 48 percent more likely to eventually reach the market, 40 percent more likely to advance from Phase 1 to Phase 2 trials, and had 32 percent fewer adverse events than drugs hitting more broadly active targets.

The researchers then pushed the system further, asking it to evaluate a protein called B7-H3 that’s associated with lung cancer. The agents discovered the protein was particularly common in connective-tissue cells called fibroblasts that are often found close to tumor cells.

The agents then discovered evidence those cells were suppressing the activity of nearby immune cells, preventing the body from detecting and reacting to the tumors. The system proposed a therapy that would tag cells expressing B7-H3 with an antibody to help direct a toxic chemotherapy drug to them.

The virtual biotech came up with its solution based solely on data available before January 2025, but in August of that year a major pharmaceutical company arrived at the same strategy independently, when its B7-H3-targeted therapy ifinatamab deruxtecan received FDA breakthrough therapy status. “This was really exciting as an independent, third-party validation that’s consistent with the effects and the design proposed by the virtual biotech,” Zou said.

However, coming up with drug targets is just one step in a long, expensive drug discovery process. While refining the candidate selection process could prevent drug companies from pursuing some obvious dead ends, it can’t speed up the rigorous lab testing and clinical trials required to get a drug to market.

Nonetheless, given the industry’s woeful record at translating promising science into finished products, an army of AI scientists that can significantly speed up a critical part of the drug discovery pipeline could be just what the doctor ordered.

The post Virtual Biotech Company Puts 37,000 AI Agents to Work on Drug Discovery appeared first on SingularityHub.