A doctor in a hospital exam room watches as a medical transcription agent updates electronic health records, prompts prescription options, and surfaces patient history in real time. A computer vision agent on a manufacturing line is running quality control at speeds no human inspector can match. Both generate non-human identities that most enterprises cannot inventory, scope, or revoke at machine speed.

That is the structural problem keeping agentic AI stuck in pilots. Not model capability. Not compute. Identity governance.

Cisco President Jeetu Patel told VentureBeat at RSAC 2026 that 85% of enterprises are running agent pilots while only 5% have reached production. That 80-point gap is a trust problem. The first questions any CISO will ask: which agents have production access to sensitive systems, and who is accountable when one acts outside its scope? IANS Research found that most businesses still lack role-based access control mature enough for today’s human identities, and agents will make it significantly harder. The 2026 IBM X-Force Threat Intelligence Index reported a 44% increase in attacks exploiting public-facing applications, driven by missing authentication controls and AI-enabled vulnerability discovery.

Why the trust gap is architectural, not just a tooling problem

Michael Dickman, SVP and GM of Cisco’s Campus Networking business, laid out a trust framework in an exclusive interview with VentureBeat that security and networking leaders rarely hear stated this plainly. Before Cisco, Dickman served as Chief Product Officer at Gigamon and SVP of Product Management at Aruba Networks.

Dickman said that the network sees what other telemetry sources miss: actual system-to-system communications rather than inferred activity. “It’s that difference of knowing versus guessing,” he said. “What the network can see are actual data communications … not, I think this system needs to talk to that system, but which systems are actually talking together.” That raw behavioral data, he added, becomes the foundation for cross-domain correlation, and without it, organizations have no reliable way to enforce agent policy at what he called “machine speed.”

The trust prerequisite that most AI strategies skip

Dickman argues that agentic AI breaks a pattern he says defined every prior technology transition: deploy for productivity first, bolt on security later.

“I don’t think trust is one of those things where the business productivity comes first, and the security is an afterthought,” Dickman told VentureBeat. “Trust actually is one of the key requirements. Just table stakes from the beginning.”

Observing data and recommending decisions carries consequences that stay contained. Execution changes everything. When agents autonomously update patient records, adjust network configurations, or process financial transactions, the blast radius of a compromised identity expands dramatically.

“Now more than ever, it’s that question of who has the right to do what,” Dickman said. “The who is now much more complicated because you have the potential in our reality of these autonomous agents.”

Dickman breaks the trust problem into four conditions. The first is secure delegation, which starts by defining what an agent is permitted to do and maintaining a clear chain of human accountability. The second is cultural readiness; he pointed to alert fatigue as a case study. The traditional fix, Dickman noted, was to aggregate alerts, so analysts see fewer items. With agents capable of evaluating every alert, that logic changes entirely.

“It is now possible for an agent to go through all alerts,” Dickman said. “You can actually start to think about different workflows in a different way. And then how does that affect the culture of the work, which is amazing.”

The third is token economics: Every agent’s action carries a real computational cost. Dickman sees hybrid architectures as the answer, where agentic AI handles reasoning while traditional deterministic tools execute actions. The fourth is human judgment. For example, his team used an AI tool to draft a product requirements document. The agent produced 60 pages of repetitive filler that immediately provided how technically responsive the architecture was, yet showed signs of needing extensive fine-tuning to make the output relevant. “There’s no substitute for the human judgment and the talent that’s needed to be dextrous with AI,” he said.

What the network sees that endpoints miss

Most enterprise data today is proprietary, internal, and fragmented across observability tools, application platforms, and security stacks. Each domain team builds its own view. None sees the full picture.

“It’s that difference of knowing versus guessing,” Dickman said. “What the network can see are actual data communications. Not ‘I think this system needs to talk to that system,’ but which systems are actually talking together.”

That telemetry grows more valuable as IoT and physical AI proliferate. Computer vision agents analyzing shopper behavior and running factory-floor quality control generate highly sensitive data that demands precise access controls.

“All of those things require that trust that we started with, because this is highly sensitive data around like who’s doing what in the shop or what’s happening on the factory floor,” Dickman said.

Why siloed agent data misses the signal

“It’s not only aggregation, but actually the creation of knowledge from the network,” Dickman said. “There are these new insights you can get when you see the real data communications. And so now it becomes what do we do first versus second versus third?”

That last question reveals where Dickman’s focus lands: the strategic challenge is sequencing, not capability.

“The real power comes from the cross-domain views. The real power comes from correlation,” Dickman said. “Versus just aggregation and deduplication of alerts, which is good, but it’s a little bit basic.”

This is where he sees the most common pitfall. Team A builds Agent A on top of Data A. Team B builds Agent B on top of Data B. Each silo produces incrementally useful automation. The cross-domain insight never materializes.

Independent practitioners validate the pattern. Kayne McGladrey, an IEEE senior member, told VentureBeat that organizations are defaulting to cloning human user profiles for agents, and permission sprawl starts on day one. Carter Rees, VP of AI at Reputation, identified the structural reason. “A significant vulnerability in enterprise AI is broken access control, where the flat authorization plane of an LLM fails to respect user permissions,” Rees told VentureBeat. Etay Maor, VP of Threat Intelligence at Cato Networks, reached the same conclusion from the adversarial side. “We need an HR view of agents,” Maor told VentureBeat at RSAC 2026. “Onboarding, monitoring, offboarding.”

Agentic AI trust gap assessment

Use this matrix to evaluate any platform or combination of platforms against the five trust gaps Dickman identified. Note that the enforcement approaches in the right column reflect Cisco’s framework.

A broken ankle and a microsegmentation lesson

Dickman grounded his framework in a scenario from his own life. A family member recently broke an ankle, which put him in a hospital exam room watching a medical transcription agent update the EHR, prompt prescription options, and surface patient history in real time. The doctor approved each decision, but the agent handled tasks that previously required manual entry across multiple systems.

The security implications hit differently when it is a loved one’s records on the screen.

“I would call it do governance slowly. But do the enforcement and implementation rapidly,” he said. “It must be done in machine speed.”

It starts with agentic IAM, where each agent is registered with defined permitted actions and a human accountable for its behavior.

“Here’s my set of agents that I’ve built. Here are the agents. By the way, here’s a human who’s accountable for those agents,” Dickman said. “So if something goes wrong, there’s a person to talk to.”

That identity layer feeds microsegmentation — a network-enforced boundary Dickman says enforces least-privileged access and limits blast radius.

“Microsegmentation guarantees that least-privileged access,” Dickman said. “You’re not relying on a bunch of host agents, which can be bypassed or have other issues.”

If the governance model works for a medical transcription agent handling patient records in an emergency department, it scales to less sensitive enterprise use cases.

Five priorities before agents reach production

1. Force cross-functional alignment now. Define what the organization expects from agentic AI across line-of-business, IT, and security leadership. Dickman sees the human coordination layer moving more slowly than the technology. That gap is the bottleneck.

2. Get IAM and PAM governance production-ready for agents. Dickman called out identity and access management and privileged access management specifically as not mature enough for agentic workloads today. Solidify the governance before scaling the agents. “That becomes the unlock of trust,” he said. “Because when the technology platform is ready, you then need the right governance and policy on top of that.”

3. Adopt a platform approach to networking infrastructure. A platform strategy enables data sharing across domains in ways fragmented point solutions cannot. That shared foundation is what makes the cross-domain correlation in the trust gap assessment above operationally real.

4. Design hybrid architectures from the start. Agentic AI handles reasoning and planning. Traditional deterministic tools execute the actions. Dickman sees this combination as the answer to token economics: it delivers the intelligence of foundation models with the efficiency and predictability of conventional software. Do not build pure-agent systems when hybrid systems cost less and fail more predictably.

5. Make the first use cases bulletproof on trust. Pick two or three high-value use cases and build them with role-based access control, privileged access management, and microsegmentation from day one. Even modest deployments delivered with best practices intact build the organizational confidence that accelerates everything after.

“You can guarantee that trust to the organization, and that will unleash the speed,” Dickman said.

That is the structural insight running through every section of this conversation. The 85% of enterprises stuck in pilot mode are not waiting for better models. They are waiting for the identity governance, the cross-domain visibility, and the policy enforcement infrastructure that makes production deployment defensible. Whether they build on Cisco’s platform or assemble their own, Dickman’s framework holds: identity governance, cross-domain visibility, policy enforcement. None of those prerequisites is optional.

The organizations that satisfy them first will deploy agents at a pace the rest cannot match, because every new agent inherits the trust architecture the first ones required. The ones still debating whether to start will watch that gap widen. Theoretical trust does not ship.

The Musk v. Altman trial entered its third week Monday, with Microsoft CEO Satya Nadella and former OpenAI co-founder and renowned AI researcher Ilya Sutskever taking the stand. Nadella testified that Elon Musk never raised concerns to him that Microsoft’s investments in OpenAI violated any special commitments, and said he viewed the partnership as clearly commercial from the start. He also described OpenAI’s 2023 board crisis as “amateur city.”

Meanwhile, Sutskever testified that he had raised concerns about Sam Altman because he feared OpenAI could be “destroyed.” He expressed concerns about Altman’s behavior to the board, in part because he said he felt “a great deal of ownership” over the startup. “I simply cared for it, and I didn’t want it to be destroyed,” Sutskever said. CNBC reports: Nadella said he was “very proud” that Microsoft took the risk to invest in OpenAI when “no one else was willing” to bet on the fledgling lab. Musk, who testified late last month, said Microsoft’s $10 billion investment was the key tipping point that made him believe OpenAI was violating its nonprofit mission. He testified that the scale of the investment bothered him, and it prompted him to open a legal investigation into OpenAI. “I was concerned they were really trying to steal the charity,” Musk said from the stand.

Nadella said he did not believe Microsoft’s investments in OpenAI were donations, and that there was a clear commercial element to their partnership from the outset. He said during the partnership’s early years, Microsoft gave OpenAI sharp discounts on computing resources, and Microsoft believed it would reap marketing benefits from doing so. During a separate video deposition that was played on Monday morning, Michael Wetter, a corporate development executive at Microsoft, said the company has recognized approximately $9.5 billion in revenue to date through its partnership with OpenAI as of March 2025.

[…] Nadella said he was “pretty surprised” by the board’s decision [to fire Altman in November 2023], and that his priority was to try and figure out how to maintain continuity for Microsoft customers. Immediately after Altman was removed, Nadella said he made an effort to learn more about what happened, adding that he suspected jealousy and poor communication was at play. During conversations with OpenAI board members after the firing, Nadella said he was simply trying to understand the language in the OpenAI’s statement about Altman being “not consistently candid” while communicating with the board. That language, Nadella said, “just didn’t sort of suffice, because this is the CEO of a company that we are invested in and we’re deeply partnered with, and so I felt that they could have explained to me what are the incidents or what is the detail behind it.” There must have been instances of jealousy or miscommunication that could have justified pushing out Altman, Nadella said. He wanted more depth from the board members after the remark about candor, but no such information was available, he said. “It was sort of amateur city, as far as I’m concerned,” Nadella testified.

[…] Musk testified that he is not entirely against OpenAI having a for-profit unit, but he said it became “the tail wagging the dog.” He repeatedly accused Altman and Brockman of enriching themselves from a charity while also reaping the positive associations that come from running a nonprofit. “Microsoft has their own motivations, and that would be different from the motivations of the charity,” Musk said from the stand. “All due respect to Microsoft, do you really want Microsoft controlling digital superintelligence?”

During a videotaped deposition shown in court last week, former OpenAI director Tasha McCauley recalled a discussion with Nadella and her fellow board members after the 2023 decision to dismiss Altman as OpenAI’s CEO. “To the best of my recollection, Satya wanted to restore things to as they had been,” McCauley said. The board members didn’t think that was the right move, she said. But as a court witness on Monday, Nadella said he never demanded that the board reinstate Altman as OpenAI CEO. Recap:

Sam Altman Had a Bad Day In Court (Day Eight)
Sam Altman’s Management Style Comes Under the Microscope At OpenAI Trial (Day Seven)
Brockman Rebuts Musk’s Take On Startup’s History, Recounts Secret Work For Tesla (Day Six)
OpenAI President Discloses His Stake In the Company Is Worth $30 Billion (Day Five)
Musk Concludes Testimony At OpenAI Trial (Day Four)
Elon Musk Says OpenAI Betrayed Him, Clashes With Company’s Attorney (Day Three)
Musk Testifies OpenAI Was Created As Nonprofit To Counter Google (Day Two)
Elon Musk and OpenAI CEO Sam Altman Head To Court (Day One)

The rapid integration of AI into healthcare devices raises ‘fresh ethical questions’, says Eoin O’Cearbhaill.

Dr Eoin O’Cearbhaill is the most recent recipient of the NovaUCD Innovation Award, a recognition given to those with success stories in commercialising research emerging from University College Dublin (UCD).

O’Cearbhaill is an associate professor in biomedical engineering at the UCD School of Mechanical and Materials Engineering, where his research focuses on developing minimally invasive medical technologies for diagnosing and treating disease.

He is also the director of the Centre for Biomedical Engineering, and leads the UCD Medical Device Design Group within the centre, which aims to address clinical needs by developing novel medical devices with real-world applications.

The research group has assisted in the creation of spin-out companies LaNua Medical, Latch Medical and Lia Eyecare, and has filed more than a dozen patents.

O’Cearbhaill has also been a member of multiple Enterprise Ireland Commercialisation Fund projects. He is a funded investigator with Research Ireland Centres Cúram and I-Form, and has consulted with a number of businesses including Boston Scientific, NeoGraft Technologies, Johnson & Johnson and CroíValve.

“A major focus of my work has been translating research from the lab into technologies that can ultimately improve patient care, including through university spin-outs and collaborations with clinicians, researchers and industry partners,” O’Cearbhaill tells SiliconRepublic.com.

What inspired you to become a researcher?

I was fortunate to grow up with parents who were always supportive of education and curiosity. Rather than one defining moment, it was a gradual realisation that research offered the opportunity to contribute something new.

Some time spent working in the medical device industry was also influential, as it gave me a practical understanding of how engineering can directly improve patient care. It showed me the importance of developing technologies that are not only innovative, but manufacturable, reliable and capable of making a real clinical impact.

Can you tell us about the research you’re currently working on?

Our research focuses broadly on minimally invasive technologies for the diagnosis and treatment of disease. Ideally, we begin by identifying an unmet clinical need, often in partnership with clinicians, and then develop practical device concepts that could address it.

My own expertise is in mechanical-based design, but coming up with the best solution relies on teamwork. We work closely with experts in electronic engineering, materials science, pharmaceuticals and clinical medicine, both internal and external to our group, to bring the right mix of skills to each challenge.

Research rarely follows a straight line. Sometimes a technology developed for one purpose reveals an unexpected opportunity elsewhere. Being able to adapt and pivot is often essential if we want to maximise the chances of clinical translation and real patient impact.

Why is your research important?

Biomedical engineering research can improve lives directly through better diagnostics, smarter treatments and less invasive procedures. It also plays an important economic role. Research attracts talented people from around the world, helps build high-value industries, and creates the environment where the next generation of Irish start-ups can emerge.

Ireland has the ingredients to become a global leader in next-generation medtech if we continue to invest in talent, translational research and entrepreneurship.

What commercial applications do you foresee for your research?

My team and I have been fortunate to work with talented researchers and entrepreneurs who have already spun technologies out of UCD with the support of NovaUCD, into companies.

These innovations span areas such as microneedle-based drug delivery, tumour treatment technologies and device-based therapies for dry eye disease and include spin-outs such as Latch Medical, LaNua Medical and Lia Eyecare.

Medical devices require robust regulation, quality systems and clinical validation, so the path to commercialisation can be challenging. But when successful, it enables research discoveries to become real products that can benefit patients at scale.

UCD Medical Device Design Group. Image: University College Dublin

What are some of the biggest challenges you face as a researcher in your field?

One of the biggest challenges is attracting and retaining outstanding talent. PhD researchers and postdoctoral fellows are central to scientific progress, so it is important that stipends and salaries remain competitive with industry, particularly during a cost-of-living crisis.

Another challenge is the rapid integration of AI and smart technologies into healthcare devices. This creates exciting new opportunities, but also raises fresh technical, regulatory and ethical questions.

Are there any common misconceptions about this area of research?

People sometimes assume that creating a medical device is only about designing the hardware and now, particularly the software in next-generation smart devices. In reality, one of the most important challenges is understanding how that device interacts with the body over time.

We are particularly interested in the interface between implantable or wearable devices and surrounding tissue. Understanding that relationship is critical if we want devices to function reliably and safely over the long term.

What are some of the areas of research you’d like to see tackled in the years ahead?

There is enormous opportunity in remote patient monitoring, technologies that enable more outpatient procedures, and devices that shorten recovery times.

These kinds of innovations can improve patient experience while also helping healthcare systems manage growing demand more efficiently.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

Bonus coupons are driving M5 MacBook Air prices lower this week, with both 13-inch and 15-inch models marked down, including configs with 24GB of RAM and 1TB of storage.

B&H’s MacBook Air sale delivers bonus discounts in the form of in-cart coupons on top of already reduced prices. With both 13-inch and 15-inch models on sale, now is a great time to pick up a thin-and-light notebook that’s was just released and is equipped with Apple’s M5 chip.

Save up to $180 on M5 MacBook Airs

• 13-inch MacBook Air (24GB, 1TB) in Sky Blue: $1,329 ($170 off)
• 15-inch MacBook Air (16GB, 512GB) in Midnight: $1,129 ($170 off)
• 15-inch MacBook Air (16GB, 1TB) in Midnight: $1,329 ($170 off)
• 15-inch MacBook Air (24GB, 1TB) in Midnight: $1,519 ($180 off)

The B&H deals above beat Amazon’s prices for the same systems, but you can find even more discounts in our 13-inch MacBook Air M5 Price Guide and 15-inch MacBook Air M5 Price Guide.

If you’d like to learn more about the M5 release, check out our hands-on M5 MacBook Air review.