Nature, Published online: 03 October 2025; doi:10.1038/d41586-025-03246-7

Researchers are increasingly turning to artificial-intelligence tools that can handle complex, multi-step processes.

Nature, Published online: 03 October 2025; doi:10.1038/d41586-025-03201-6

Machine learning can speed up the discovery of potential antibiotics, but challenges remain.

Nature Medicine, Published online: 30 September 2025; doi:10.1038/s41591-025-03933-y

In a prospective study enrolling 1,222 patients from 22 emergency departments, a device using a machine-learning-based signature of blood mRNAs demonstrated clinically acceptable performance to diagnose bacterial and viral infections and to predict the all-cause need for critical care interventions within 7 days, with benchmark to established biomarkers and risk scores.

Guest author: Or Hillel, Green Lamp

Applications have become the foundation of how organisations deliver services, connect with customers, and manage important operations. Every transaction, interaction, and workflow runs on a web app, mobile interface, or API. That central role has made applications one of the most attractive and frequently-targeted points of entry for attackers.

As software grows more complex, spanning microservices, third-party libraries, and AI-powered functionality, so do the security risks. Traditional scanning methods struggle to keep up with rapid release cycles and distributed architectures. This has opened the door for AI-driven application security tools, which bring automation, pattern recognition, and predictive capabilities to a field that once relied heavily on manual reviews and static checks.

Best practices for using AI AppSec tools

To get the most value from AI-powered application security, teams should follow some key best practices:

1. Shift security left: Integrate tools early in the SDLC so issues are caught before production.
2. Combine approaches: Use AI tools alongside traditional SAST, DAST, and manual reviews to cover all bases.
3. Enable continuous learning: Choose solutions that improve over time by ingesting threat intelligence and user feedback.
4. Keep humans in the loop: AI should augment, not replace, human judgment. Security experts are still needed for complex decision-making.
5. Align with compliance: Ensure AI-powered findings can be mapped to regulatory requirements like SOC 2, HIPAA, or GDPR.

The 5 best AI-powered AppSec tools of 2025

1. Apiiro

Apiiro is reinventing the way organisations assess and manage risk in the modern software supply chain. It moves beyond legacy scanning to implement true risk intelligence, offering full-stack, contextual analysis powered by deep AI.

Apiiro brings visibility not only to what vulnerabilities exist in code and dependencies, but also to how changes, developer actions, and business context interact to shape risk. Its AI systems process data from source control, CI/CD pipelines, cloud configurations, and user access patterns, allowing it to prioritise remediation based on business impact.

2. Mend.io

Mend.io has rapidly evolved into a cornerstone of the AI-driven AppSec ecosystem, addressing the full spectrum of risks facing software teams today. Using machine learning and advanced analytics, Mend.io is purpose-built to handle the security challenges of code produced by both humans and artificial intelligence.

Leading organisations are attracted to Mend.io’s unified platform, which delivers seamless coverage for source code, open source, containers, and AI-generated functional logic. Its capabilities extend far beyond detection, enabling rapid, automated, and context-rich remediation that saves engineering time and reduces business exposure.

3. Burp Suite

Burp Suite has long been a foundational tool for web application security professionals, but its latest AI-driven evolution makes it essential for defending cutting-edge app landscapes. Today, Burp Suite combines traditional manual penetration testing strengths with sophisticated machine learning, delivering smarter scanning and deeper insight than ever before.

Where legacy DAST (Dynamic Application Security Testing) tools might struggle with modern, dynamic, or API-rich applications, Burp Suite’s AI modules adapt to changes in real time, learning from traffic patterns and user behaviours to uncover anomalies and hard-to-spot vulnerabilities.

4. PentestGPT

PentestGPT represents the future of automated offensive security, using generative AI to simulate the tactics of contemporary adversaries. Unlike pattern-based scanners, PentestGPT can devise new attack paths, generate custom payloads, and think creatively about bypassing controls and protections.

PentestGPT blends autonomous testing with educational support: security analysts, testers, and developers can interact with the platform conversationally, gaining hands-on guidance for complex scenarios and real-world exploit development.

5. Garak

Garak is an emerging leader specialising in security for AI-driven applications, specifically, large language models, generative agents, and their integration into wider software systems. As organisations increasingly embed AI into customer interactions, business logic, and automation, new risks have arisen that traditional AppSec tools simply weren’t built to address.

Garak is designed to probe and harden these AI-infused interfaces, ensuring models respond safely and preventing AI-specific exploits like prompt injections and privacy breaches.

Core features of AI-driven AppSec tools

While not every solution offers the same features, most AI-powered application security tools share several core capabilities:

1. Intelligent vulnerability detection

AI models trained on massive datasets of known exploits can spot coding errors, misconfigurations, and insecure dependencies more accurately than static rule-based tools. They adapt over time, improving detection with each new dataset.

2. Automated remediation guidance

One of the major pain points in AppSec is not just finding vulnerabilities but knowing how to fix them. AI tools can generate remediation advice tailored to the specific context, often offering code suggestions or step-by-step fixes.

3. Continuous monitoring and real-time analysis

Instead of one-time scans, AI-powered tools continuously monitor applications in production. They analyse runtime behaviour, API calls, and data flows to spot anomalies that could indicate an active attack.

4. Risk prioritisation

AI can evaluate the severity of each vulnerability based on exploitability, business impact, and external threat intelligence. The ensures that teams focus on the issues most likely to cause real damage.

5. Integration with DevOps workflows

Modern AppSec tools embed directly into CI/CD pipelines, issue trackers, and developer environments. AI accelerates these processes by automating tasks that previously slowed down builds or required manual oversight.

Building resilient software in an AI world

AI-powered application security is not a single tool, process, or department, it’s the foundation on which resilient, innovative, and trusted software is built. In 2025, the leaders in this space are not just those who scan for vulnerabilities, but those who can learn, adapt, and protect at the velocity of AI-driven innovation.

From comprehensive risk intelligence and agile remediation to the defense of AI-generated code and AI agents themselves, today’s AppSec solutions are reshaping what’s possible, and what’s necessary, for digital security in any industry.

Guest author: Or Hillel, Green Lamp

The post The 5 best AI AppSec tools in 2025 appeared first on AI News.

Nature, Published online: 01 October 2025; doi:10.1038/d41586-025-03147-9

Critics slammed attempts by Google, Microsoft and Meta to speed up materials discovery. But behind the hype, there is progress.

Nature, Published online: 01 October 2025; doi:10.1038/d41586-025-02954-4

How humans behave.

A team at Microsoft says it used artificial intelligence to discover a “zero day” vulnerability in the biosecurity systems used to prevent the misuse of DNA.

These screening systems are designed to stop people from purchasing genetic sequences that could be used to create deadly toxins or pathogens. But now researchers led by Microsoft’s chief scientist, Eric Horvitz, say they have figured out how to bypass the protections in a way previously unknown to defenders. 

The team described its work today in the journal Science.

Horvitz and his team focused on generative AI algorithms that propose new protein shapes. These types of programs are already fueling the hunt for new drugs at well-funded startups like Generate Biomedicines and Isomorphic Labs, a spinout of Google. 

The problem is that such systems are potentially “dual use.” They can use their training sets to generate both beneficial molecules and harmful ones.

Microsoft says it began a “red-teaming” test of AI’s dual-use potential in 2023 in order to determine whether “adversarial AI protein design” could help bioterrorists manufacture harmful proteins. 

The safeguard that Microsoft attacked is what’s known as biosecurity screening software. To manufacture a protein, researchers typically need to order a corresponding DNA sequence from a commercial vendor, which they can then install in a cell. Those vendors use screening software to compare incoming orders with known toxins or pathogens. A close match will set off an alert.

To design its attack, Microsoft used several generative protein models (including its own, called EvoDiff) to redesign toxins—changing their structure in a way that let them slip past screening software but was predicted to keep their deadly function intact.

The researchers say the exercise was entirely digital and they never produced any toxic proteins. That was to avoid any perception that the company was developing bioweapons. 

Before publishing the results, Microsoft says, it alerted the US government and software makers, who’ve already patched their systems, although some AI-designed molecules can still escape detection. 

“The patch is incomplete, and the state of the art is changing. But this isn’t a one-and-done thing. It’s the start of even more testing,” says Adam Clore, director of technology R&D at Integrated DNA Technologies, a large manufacturer of DNA, who is a coauthor on the Microsoft report. “We’re in something of an arms race.”

To make sure nobody misuses the research, the researchers say, they’re not disclosing some of their code and didn’t reveal what toxic proteins they asked the AI to redesign. However, some dangerous proteins are well known, like ricin—a poison found in castor beans—and the infectious prions that are the cause of mad-cow disease.

“This finding, combined with rapid advances in AI-enabled biological modeling, demonstrates the clear and urgent need for enhanced nucleic acid synthesis screening procedures coupled with a reliable enforcement and verification mechanism,” says Dean Ball, a fellow at the Foundation for American Innovation, a think tank in San Francisco.

Ball notes that the US government already considers screening of DNA orders a key line of security. Last May, in an executive order on biological research safety, President Trump called for an overall revamp of that system, although so far the White House hasn’t released new recommendations.

Others doubt that commercial DNA synthesis is the best point of defense against bad actors. Michael Cohen, an AI-safety researcher at the University of California, Berkeley, believes there will always be ways to disguise sequences and that Microsoft could have made its test harder.

“The challenge appears weak, and their patched tools fail a lot,” says Cohen. “There seems to be an unwillingness to admit that sometime soon, we’re going to have to retreat from this supposed choke point, so we should start looking around for ground that we can actually hold.” 

Cohen says biosecurity should probably be built into the AI systems themselves—either directly or via controls over what information they give. 

But Clore says monitoring gene synthesis is still a practical approach to detecting biothreats, since the manufacture of DNA in the US is dominated by a few companies that work closely with the government. By contrast, the technology used to build and train AI models is more widespread. “You can’t put that genie back in the bottle,” says Clore. “If you have the resources to try to trick us into making a DNA sequence, you can probably train a large language model.”

Background: Development and use of digital public health tools surged during the COVID-19 pandemic. Among these tools, vaccine verification systems emerged as alternatives to paper vaccine records, aiming to help limit the spread of disease. In November 2021, the Washington State Department of Health launched “WA Verify,” a QR code–based vaccine verification system built on the SMART Health Card framework, providing residents with a convenient way to store and share proof of vaccination digitally. However, WA Verify was developed and deployed before assessments and public input regarding potential adoption challenges—such as concerns about privacy, surveillance, data sharing, trust in the technology, and the managing organizations—could be completed. Objective: This analysis used statewide survey data from Washington to identify and characterize barriers and facilitators to the adoption of WA Verify, and to understand how factors such as data privacy, security, attitudes toward public health policies and communication, and technological proficiency may influence acceptance and uptake of digital public health tools. Methods: A cross-sectional statewide survey was distributed between September 2022 and January 2023 to a random sample of 5000 Washington households. Respondents were categorized into 3 groups based on their responses indicating WA Verify “users,” “potential users,” or “unlikely users.” Comparisons were made between groups regarding experiences with and opinions on COVID-19 vaccine and test verification, public health policies, communication, digital tools, technological proficiency, sociodemographic characteristics, and health history. Poststratification weights were applied to reduce nonresponse bias. Results: Of the 1401 respondents, 359 (25.6% unweighted, 25.8% weighted) were users, 662 (47.3% unweighted, 49.8% weighted) were potential users, and 380 (27.1% unweighted, 24.4% weighted) were unlikely users. All percentages reported are based on weighted data. Compared with users and potential users, unlikely users were more likely to oppose policies requiring proof of COVID-19 vaccination or negative test results (users: 6.0%, potential users: 13.6%, unlikely users: 65.9%). Unlikely users were more likely to cite concerns about personal health data security and phone hacking or tracking, though these concerns were also notable among potential users and users. Users and potential users were more likely to perceive a digital vaccine verification system as convenient (users: 96.5%, potential users: 92.3%, unlikely users: 38.1%) and indicated openness to receiving relevant information from a range of sources. Unlikely users were more likely to report not owning a smartphone and demonstrated lower technological proficiency (users: 12.3%, potential users: 15.9%, unlikely users: 32.3%), indicating a technological divide between groups. Conclusions: While nearly three-quarters of respondents had either already adopted or were willing to adopt a tool like WA Verify, concerns about data security, lower technological proficiency, and distrust of public health characterized those least likely to adopt such tools. Identifying barriers to adoption among “unlikely users” is essential for developing effective communication strategies—such as targeted marketing and community engagement—to improve adoption and ensure equitable access to public health technologies.

Nature Biotechnology, Published online: 01 October 2025; doi:10.1038/s41587-025-02844-0

A microchip-based DNA synthesis method enables scalable production of complex DNA constructs.

On Wednesday, Oct. 1, patients across America lost access to care they had the day before not because medicine changed, but because politics did.

When the government shut down, so did federal telemedicine flexibilities tied to pandemic-era waivers. These waivers allowed Medicare and many private insurers to reimburse virtual visits at parity with in-person care, across settings and sometimes across state lines. With the waivers lapsed, reimbursement is uncertain and in some cases discontinued.

Read the rest…

© Adobe

npj Digital Medicine, Published online: 01 October 2025; doi:10.1038/s41746-025-02004-3

Large language models forecast patient health trajectories enabling digital twins