logo
Today's AI Could Make Pandemics 5 Times More Likely

Today's AI Could Make Pandemics 5 Times More Likely

Recent developments in AI could mean that human-caused pandemics are five times more likely than they were just a year ago, according to a study of top experts' predictions shared exclusively with TIME.
The data echoes concerns raised by AI companies OpenAI and Anthropic in recent months, both of which have warned that today's AI tools are reaching the ability to meaningfully assist bad actors attempting to create bioweapons.
Read More: Exclusive: New Claude Model Triggers Bio-Risk Safeguards at Anthropic
It has long been possible for biologists to modify viruses using laboratory technology. The new development is the ability for chatbots—like ChatGPT or Claude—to give accurate troubleshooting advice to amateur biologists trying to create a deadly bioweapon in a lab. Safety experts have long viewed the difficulty of this troubleshooting process as a significant bottleneck on the ability of terrorist groups to create a bioweapon, says Seth Donoughe, a co-author of the study. Now, he says, thanks to AI, the expertise necessary to intentionally cause a new pandemic 'could become accessible to many, many more people.'
Between December 2024 and February 2025, the Forecasting Research Institute asked 46 biosecurity experts and 22 'superforecasters' (individuals with a high success rate at predicting future events) to estimate the risk of a human-caused pandemic. The average survey respondent predicted the risk of that happening in any given year was 0.3%.
Crucially, the surveyors then asked another question: how much would that risk increase if AI tools could match the performance of a team of experts on a difficult virology troubleshooting test? If AI could do that, the average expert said, then the annual risk would jump to 1.5%—a fivefold increase.
What the forecasters didn't know was that Donoughe, a research scientist at the pandemic prevention nonprofit SecureBio, was testing AI systems for that very capability. In April, Donoughe's team revealed the results of those tests: today's top AI systems can outperform PhD-level virologists at a difficult troubleshooting test.
Read More: Exclusive: AI Outsmarts Virus Experts in the Lab, Raising Biohazard Fears
In other words, AI can now do the very thing that forecasters warned would increase the risk of a human-caused pandemic fivefold. (The Forecasting Research Institute plans to re-survey the same experts in future to track whether their view of the risks has increased as they said it would, but said this research would take months to complete.)
To be sure, there are a couple of reasons to be skeptical of the results. Forecasting is an inexact science, and it is especially difficult to accurately predict the likelihood of very rare events. Forecasters in the study also revealed a lack of understanding of the rate of AI progress. (For example, when asked, most said they did not expect AI to surpass human performance at the virology test until after 2030, while Donoughe's test showed that bar had already been met.) But even if the numbers themselves are taken with a pinch of salt, the authors of the paper argue, the results as a whole still point in an ominous direction. 'It does seem that near-term AI capabilities could meaningfully increase the risk of a human-caused epidemic,' says Josh Rosenberg, CEO of the Forecasting Research Institute.
The study also identified ways of reducing the bioweapon risks posed by AI. Those mitigations broadly fell into two categories.
The first category is safeguards at the model level. In interviews, researchers welcomed efforts by companies like OpenAI and Anthropic to prevent their AIs from responding to prompts aimed at building a bioweapon. The paper also identifies restricting the proliferation of 'open-weights' models, and adding protections against models being jailbroken, as likely to reduce the risk of AI being used to start a pandemic.
The second category of safeguards involves imposing restrictions on companies that synthesize nucleic acids. Currently, it is possible to send one of these companies a genetic code, and be delivered biological materials corresponding to that code. Today, these companies are not obliged by law to screen the genetic codes they receive before synthesizing them. That's potentially dangerous because these synthesized genetic materials could be used to create mail-order pathogens. The authors of the paper recommend labs screen their genetic sequences to check them for harmfulness, and for labs to implement 'know your customer' procedures.
Taken together, all these safeguards—if implemented—could bring the risk of an AI-enabled pandemic back down to 0.4%, the average forecaster said. (Only slightly higher than the 0.3% baseline of where they believed the world was before they knew today's AI could help create a bioweapon.)
'Generally, it seems like this is a new risk area worth paying attention to,' Rosenberg says. 'But there are good policy responses to it.'
Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

What To Expect From the Magnificent Seven in the Second Half of 2025
What To Expect From the Magnificent Seven in the Second Half of 2025

Yahoo

time12 minutes ago

  • Yahoo

What To Expect From the Magnificent Seven in the Second Half of 2025

Analysts expect the group of tech giants to continue to benefit from their size and position in the AI race. They also warn that their earnings growth relative to other leading companies may slow. And even in AI, analysts warn, investors may start to look to other stocks in search of gains. Three of the Mag 7—Nvidia, Microsoft, and Meta—are up double digits since the start of 2025 and are currently trading at or near record Magnificent Seven entered 2025 on a high note. Since then, the tune has meandered all over the place. Looking ahead, analysts expect the group of tech giants to continue to benefit from their size and position in the AI race, which could both fuel future growth and offer protection for investors concerned about trade-fueled uncertainty. But they also warn that their earnings growth relative to other leading companies may slow—and even in artificial intelligence, investors may start to look to other stocks in search of gains. Below, we'll catch you up on the year so far for the Magnificent Seven—and go into more detail about some of the likely drivers of their performance that await in the months to come. xExcitement about AI propelled the tech giants—Apple (AAPL), Microsoft (MSFT), Nvidia (NVDA), Amazon (AMZN), Alphabet (GOOG), Meta (META), and Tesla (TSLA)—to two years of outsized gains. The stocks, like the broader market, were pushed higher by post-election optimism about President-elect Donald Trump's promises to cut taxes, roll back regulations, and welcome the business community to Washington with wide-open arms. No company stood to benefit more than Tesla, whose CEO Elon Musk was expected to wield immense influence within the White House after publicly, and expensively, supporting Trump's campaign. Instead, Tesla's sales–and stock–crashed as Musk took a public role in Trump's administration that led to both political opposition and concern about his work with the carmaker. Meanwhile, Trump's tariffs sparked panic on Wall Street that pummeled high-flying tech stocks. By the time Trump paused the tariffs, the Roundhill Magnificent Seven ETF (MAGS) was trading more than 30% below its December high. Things have recovered since. Easing trade tensions, a strong U.S. economy, and resilient businesses helped the 'Mag 7' claw back nearly all of those losses in the second quarter, with the Roundhill ETF having edged into the green year-to-date. Three of the Mag 7—Nvidia, Microsoft, and Meta—are up double digits since the start of 2025 and are currently at or near record highs. Amazon and Alphabet remain slightly off their records. Apple and Tesla are down 14% and 19%, respectively, year-to-date. These tech titans face plenty of risks—including high valuations, ongoing tariff negotiations, and geopolitical tensions that could threaten their businesses—in the second half. But experts say they also have the opportunity to use their size and deep pockets to bolster their positions in AI, which could lead to both long-term gains and near-term share-price benefits. At times in the first half of 2025, it looked like tech giants might scale back their AI investments. The success of China's DeepSeek and its efficient AI reasoning model raised questions about whether hyperscalers needed to add as much computing capacity as expected. Trump's implementation of sweeping tariffs threatened to plunge the U.S. into a period of stagflation and suppress consumer and business spending. Hyperscalers stood by plans to continue spending big on AI. Microsoft, Amazon, Alphabet, and Meta this year all indicated that their cloud and AI businesses were constrained by insufficient computing capacity. Cumulatively, the four companies are expected to spend more than $300 billion on infrastructure in 2025, with much of that earmarked for data centers and equipment required to train and deploy AI. That spending is expected to continue benefitting the companies that design, make, and market the most advanced semiconductors, including Nvidia and Broadcom (AVGO). It should also boost sales of networking technology companies like Arista Networks (ANET), Amphenol (APH), and Coherent (COH). The Mag 7 have been the main drivers of S&P 500 earnings growth in the last two years. The group's profits grew nearly 28% in the first quarter, slightly below their average over the prior three quarters. The remainder of the S&P 500 reported growth of about 9%. The gap between the two groups, now 19 percentage points, was nearly 30 percentage points as recently as the second quarter of 2024. That gap is expected to narrow further over the next year, with FactSet projecting the rest of the index's growth will be on par with the Mag 7's by the first quarter of 2026. A possible caveat: Over the past year, analysts have consistently overestimated how quickly the broader market would catch up with the Mag 7. Tariffs and economic uncertainty could help the Magnificent 7 in the second half. Analysts at Janus Henderson expect second-quarter U.S. earnings, which kick off with big bank results in mid-July, will come under pressure from tariff anxiety before rebounding later in the year as the trade outlook becomes clearer and mitigation strategies take effect. 'Companies with strong balance sheets, scale, pricing power, and supply chain flexibility could weather this earnings pressure and recover faster,' they wrote. Most of the Mag 7 operate high-margin businesses. All have scale that should give them a competitive advantage in times of uncertainty. Against 'a backdrop of sluggish interim growth and higher-for-longer rate environment, we are likely to see a repeat of the 2023-2024 playbook of unhealthy narrow market leadership and high market concentration,' JPMorgan analysts expect. The extent to which the Mag 7 companies are synonymous with the AI trade could decline and take some of the wind out of their stocks' sails. JP Morgan analysts expect 'a broadening AI theme' that could 'accelerate further with the potential for greater productivity and efficiency gains.' Semiconductor, power, data center, and cybersecurity are their preferred AI themes outside the Mag 7. To be sure, the Mag 7 are still some of Wall Street's favorite AI stocks. 'Our preferred way to play the AI theme are the hyperscalers,' particularly Microsoft, 'and key data/analytics consumption names,' including Snowflake (SNOW) and MongoDB (MDB), said Citibank application software analyst Tyler Radtke. Citi analysts covering systems and back-office software have also emphasized the importance of AI monetization in the coming months. Companies that can develop AI programs that improve their customers' efficiency—like Cyberark (CYBR) in the cybersecurity space and (MNDY) in project management software—are best positioned to lead the AI rally, some argue. Read the original article on Investopedia Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data

Companies are relying on aptitude and personality tests more to combat AI-powered job hunters
Companies are relying on aptitude and personality tests more to combat AI-powered job hunters

Business Insider

time16 minutes ago

  • Business Insider

Companies are relying on aptitude and personality tests more to combat AI-powered job hunters

Are you happy? Do you sleep well? Do you have many friends? Are you a workaholic? Those are some of the questions Katelin Eagan, 27, said she had to answer recently when she was applying for a job. She agreed to take a cognitive and personality assessment as part of the hiring process, but was a bit bewildered. Many of the questions had nothing to do with the engineering position, which, after completing the tests and going through several months of silence, she was eventually rejected for. Eagan says she's been applying for jobs full-time since the start of the year. Her efforts haven't panned out yet, which she attributes partly to how competitive her field has become and employers having room to be picky. "I think there's definitely a lower amount than I thought there would be," she said of available roles. But that may be only part of the story. Employers are growing increasingly selective, partly because many are seeing a flood of seemingly perfect candidates, many of whom are suspected of using AI to finesse their applications, according to recruiters and hiring assessment providers who spoke to BI. The solution many companies have come to? Make everyone take a test — and see who candidates really are, irrespective of what ChatGPT suggested they put on their résumés. According to surveys conducted by TestGorilla, one firm that administers talent assessments for employers, 76% of companies that had hired in the 12 months leading up to April said they were using skills tests to determine if a candidate was a right fit, up from 55% who said they were using role-specific skills tests in 2022. Employers seem most interested in testing for soft skills — amorphous qualities like communicativeness and leadership — as well as administering general aptitude and personality tests, Wouter Durville, the CEO of TestGorilla, told Business Insider. TestGorilla's Critical Thinking test was completed more than 100,000 times in the first quarter of this year, a 61% increase compared to the same quarter in 2024. The firm also offers a Big 5 personality assessment, which was completed more than 127,000 times in the first quarter — a 69% increase compared to last year. Demand among US employers in particular has been "massive," Durville said, adding that many firms have turned to tests as a result of being overwhelmed with job applications. The US is the largest market for the firm, which is based in the Netherlands. "The biggest thing is people just want to hire the best people. It's very selfish and it's fine," Durville said. Canditech, another firm that offers hiring assessments, says it's also seen rapid growth in the last year. In 2024, the assessment usage grew 135% compared to the prior year, CEO Guy Barel told BI. He estimates that assessment usage is on track to soar 242% year-over-year. Barel says the surge is partly due to the job market tipping more in favor of employers. In many cases, companies he works with are flooded with "tons of candidates" and looking to "move forward as fast as possible," he said. Criteria, another skills-based assessment provider, says test usage has more than doubled in recent years. "AI is kind of creating this authenticity crisis in talent acquisition, because everyone can and is putting their résumé into ChatGPT." Criteria CEO Josh Millet told BI. "It's all about demonstrating your ability or your skill or your personality in an objective way that's a little bit harder to fake." The AI job market Jeff Hyman, a veteran recruiter and the CEO of Recruit Rockstars, estimates that demand for testing among his clients has increased by around 50% over the last 18 months. That's due to a handful of different reasons, he said — but companies being inundated by job applications is near the top, thanks to candidates leaning more on AI to gain an edge and send out résumés en masse, he says. Hyman says a typical job he tries to fill for a client has around 300 to 500 applicants, though he's spoken to companies trying to fill roles with more than 1,000 candidates within several days of being posted online. The number of job applications in the US grew at more than four times the pace of job requisitions in the first half of 2024, according to a report from WorkDay. Companies also want to test candidates' soft skills as remote work grows more common, Hyman adds — and they want to be sure they're getting the right person. Depending on the size of the organization, a bad hire can cost a company anywhere from $11,000 to $24,000, a survey conducted by CareerBuilder in 2016 found. According to TestGorilla, 69% of employers who issued tests this year said they were interested in assessing soft skills, while 50% said they were interested in assessing a candidate's cognitive ability. A separate survey by Criteria ranked emotional intelligence as the most sought-after skill among employers, followed by analytical thinking. "It's about their personality and to see if they are a good fit to the organization, if they share the same DNA," Durville said, though he noted that, in many cases, companies find the results of the tests to be shaky as a sole evaluation metric. TestGorilla, Canditech, and Criteria told BI that employers say they're enjoying the time and cost savings of administering tests. According to TestGorilla, 82% of employers who said they used skills-based hiring — a catch-all term for hiring based on proven skills — said they were satisfied with new hires, compared to 73% of US employers on average. Canditech, meanwhile, claims its assessments can help employers cut down on hiring time by as much as 50%, and reduce "unnecessary interviews" by as much as 80%, according to its website. But Hyman thinks there are some issues with hiring tests. For one, he says employers turn down candidates who don't score well "all the time," despite them being otherwise qualified for the job. The trend also appears to be turning off job candidates. Hyman estimates around 10%-20% of applicants will outright refuse to take a test if employers introduce it as a first step in the hiring process, though that's a practice Canditech's Barel says is becoming increasingly common. Hyman says he frequently has conversations with employers urging them not to put so much weight on test results, due to the potential for a mis-hire. "That's lazy hiring, to be honest. I think that's not the right way to go about it," he said.

I'm a CEO running an 8-figure AI company. I'm also an extreme procrastinator — and I think that's a good thing.
I'm a CEO running an 8-figure AI company. I'm also an extreme procrastinator — and I think that's a good thing.

Business Insider

time18 minutes ago

  • Business Insider

I'm a CEO running an 8-figure AI company. I'm also an extreme procrastinator — and I think that's a good thing.

This as-told-to essay is based on a transcribed conversation with Richard White, CEO of AI note-taking company Fathom. The following has been edited for length and clarity. Everyone talks about procrastination as a personal failing. I disagree. I'm an extreme procrastinator, and I've been building successful companies, like UserVoice and, most recently, Fathom, for 15 years. It's been one of my greatest assets as an entrepreneur. I see procrastination as ruthless prioritization in disguise. Consider procrastination as data collection Procrastination is a way to gather more information before making critical decisions. When I delay a choice, I'm not being lazy; I'm waiting for the optimal moment when I have enough data to make the right call. In college, I judged the size of a project and left it to the last achievable minute. I might have frustrated my peers or not gotten the most out of every seminar, but I'd do exactly what was needed and nothing more. Since then, I've learned to be more thoughtful about my approach. I used this philosophy to build Fathom, which now has an eight-figure valuation. We started building the company in 2020. Instead of rushing to market with whatever technology was available, we waited. We gathered data. We watched AI capabilities evolve. For example, prior to the rollout of GPT-4 and Claude 2, Fathom would yield basic call summaries. When GPT-4 was made available, we saw its capabilities and knew concerted investment on our side would yield massive gains. It was a foundation for our more advanced call summary features, and any earlier investment wouldn't have been as useful to our company. The same principle applies to my personal life. I plan trips at the last minute because I want to see what opportunities emerge, what's actually happening in my life, and what I might miss out on if I commit too early. In other work environments or even relationships, being a procrastinator can annoy people. However, the real and most common downside of procrastinating is underestimating the effort required and starting something too late to meet the deadline. As a CEO, I get to define the deadlines or, in our case, create a deadline-free environment. Urgent matters to trump important matters I've adopted an unfashionable approach for a CEO: urgent trumps important. This keeps our entire company moving forward without anyone waiting on me to make progress. It means that sometimes important but non-urgent things languish. I tell my team that if something's truly important, they should keep tagging me until I respond. This creates a culture where people at all levels in the company can advocate for what matters, and truly important tasks don't get lost. I've developed what I call the "Jenga model" for running my company. Like the game, when a piece looks too difficult or risky to move, I leave it and come back to it later. I can think about a problem and then put it back down without fear. Months later, I'll pick it up again, and suddenly, the answer falls right out. I'll prioritize problems that will get bigger with time, such as making an important product change, as well as problems where the solutions are low stakes or reversible. Higher-stakes decisions that are non-reversible should be deferred to gather data as long as possible, or broken out into lower-stakes decisions that help gather data to inform the larger issue. For product development, we circulate ideas internally while waiting for technological improvements. We don't rush features to market. Instead, we wait for the AI to get better, watch for what could go wrong, and optimize our timing. I don't think I have ever missed out on an opportunity. The reality in startups is that few things have a "hard" deadline. Implementing a deadline-free environment at Fathom means there hasn't been much negative feedback on this model. My team understands what we're prioritizing versus what we're doing later. CEOs need to play to their strengths Working alongside great entrepreneurs over the years has taught me that you can't build something around yourself that doesn't play to your strengths. My strength isn't planning or rigid schedules. My strength is recognizing optimal timing, gathering information, and making high-impact decisions. I delegate open-ended goals to my teams rather than micromanaging tasks. I encourage people at every level to make decisions. Most people think efficiency means doing things as quickly as possible. I think efficiency means doing things at the right time. You might be wrong about when something is needed or the time cost of execution, but that's the risk you take using your best collective judgment. This mindset has served Fathom incredibly well. We're exploring ways to use AI to take better notes, reduce unnecessary meetings, and democratize information sharing within companies. The next time someone tells you that procrastination is holding you back, ask yourself: Are you really procrastinating, or are you waiting for better information? Are you being lazy, or are you being strategically patient? Sometimes the best thing you can do is put the problem down and come back to it when you can solve it easily and effectively.

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into a world of global content with local flavor? Download Daily8 app today from your preferred app store and start exploring.
app-storeplay-store