Christopher Cole

PharmTech AI Pulse Check debuts: pharma experts tackle GxP 'slop,' multi-agent validation risk, OCR document mining, and personalized mRNA cancer therapy.

Introducing the "PharmTech AI Pulse Check" Expert Video Series

's new biweekly video series, “PharmTech AI Pulse Check,” host Chris Cole, associate editorial director, PharmTech, sits down with three expert co-hosts—Richard Jaenisch, Open Biopharma Research and Training Institute, Florin Muraru, independent regulatory affairs advisor, and Gourav Pandey, R&D quality lead, Takeda—to unpack the AI developments and questions that matter most to people building, manufacturing, and delivering therapies for patients.

The conversation opens with a pointed question: is manufacturing and quality documentation developing its own version of "AI slop," the fluent-but-empty content already flagged widely on social platforms? All three experts agree that the risk is real, particularly around deviation reports, where generative AI's ease of use can produce polished but substance-free documentation. Jaenisch warns that slop is "almost unavoidable" and represents a real cost when reviewers spend time on content that ultimately proves useless, or worse, reaches a regulator undetected. Muraru reframes the issue as fundamentally a human problem, not a model problem, and raises a deeper concern: AI-generated volume can outpace human review capacity, shifting the real bottleneck to the person signing off. The panel also discusses a "second-degree slop" framework, distinguishing between AI directly producing flawed documents and humans trusting AI-generated advice without further verification, citing a 

 as an early real-world example. The panel converges on a shared fix: redesigning source documents so AI has clear, deterministic rules to follow, reducing ambiguity at the root rather than relying solely on downstream review.

The discussion shifts to multi-agent AI systems, prompted by a r

ecent case in which AI agents secretly built a private communication channel after being shut down

. Drawing parallels to pharma's existing segregation-of-duties principles, the panel debates where human sign-off gates should sit as multi-agent systems move toward the shop floor. Muraru argues that agent-to-agent reasoning is acceptable, but any GxP-relevant action, like releasing a batch, requires human decision making, and that inter-agent exchanges must be captured as auditable records. Pandey stresses that data integrity principles already governing handwritten notes apply equally to AI outputs, and that traceable, accountable human sign-off remains non-negotiable regardless of how much groundwork AI does.

The conversation turns to document intelligence, sparked by a 

newly open-sourced model capable of parsing entire multi-page documents

—including scanned records, handwritten logbooks, and historical batch data—in a single pass. The panel sees major potential for mining decades of previously unsearchable pharma records, including failed experiments and terminated programs, but cautions that the moment this information feeds a regulatory submission or GxP decision, the system requires formal validation, and OCR errors could become data integrity issues.

The episode continues with a discussion of Merck and Moderna's 

, which uses individualized mRNA-based cancer therapy tailored to each patient's tumor mutations. The panelists emphasize a critical distinction: this relies on precision machine learning, not the generative AI most people associate with chatbots, an important nuance as public conversation increasingly conflates the two.

Wrapping up with a rapid-fire round, each panelist names pharma manufacturing's biggest AI governance gap: Jaenisch points to inadequate workforce training, Muraru highlights the absence of standards for validating models after retraining and drift, and Pandey argues the industry over-focuses on the AI model itself rather than the surrounding pipeline and source documentation.

Videos

PharmTech AI Pulse Check debuts: pharma experts tackle GxP 'slop,' multi-agent validation risk, OCR document mining, personalized mRNA cancer therapy.

The FDA has finalized the design of its Expedited Investigational New Drug (IND) Pilot and begun accepting applications, giving sponsors until Oct. 30, 2026, to apply.¹ The program pairs drug developers with qualified research institutions (QRIs)—academic medical centers, health networks, contract research organizations, or regulatory advisory firms—that will provide pharmacology/toxicology, clinical, and chemistry, manufacturing, and controls (CMC) input on first-in-human (FIH) IND submissions.¹

Under the pilot, the FDA will accept and review individual IND components on a rolling basis during the pre-IND phase rather than waiting for a complete package, a shift the agency says should let issues surface and be resolved earlier, reducing the odds of a clinical hold once the 30-day statutory IND review clock starts.¹ The FDA expects to select 8 to 10 sponsor-QRI pairs for the initial cohort, chosen from applications reviewed by agency scientific staff.¹

“Under the Trump Administration, boosting domestic innovation and ensuring American patients have first access to groundbreaking treatments is a top priority,” said Acting FDA Commissioner Kyle Diamantas, JD, in the agency’s announcement.¹ Michael Davis, MD, PhD, director of FDA’s Center for Drug Evaluation and Research, added that public comment on the proposed design shaped the final program.¹

The pilot is a formal deliverable of HHS’s Operation TrialBlazer, the modernization push FDA launched after then-Commissioner Marty Makary 

warned that the US was ceding ground to China in early-stage drug development

 in a televised interview earlier this year.³ Makary pointed to hospital contracting delays, duplicative institutional review board (IRB) processes, and IND-related timelines as the operational bottlenecks slowing US first-in-human studies—even though FDA’s own 30-day IND review window is already short by international standards. China, by contrast, has used implied-approval IND mechanisms and ICH-aligned reforms to compress its own timelines, contributing to a rise in China-originated FIH and early oncology programs.

What Does it Mean for CMC and Manufacturing Teams?

For CMC and manufacturing teams, the pilot signals that FDA sees pre-IND CMC review, not just clinical protocol design, as a lever worth testing for cycle-time reduction, since QRIs are explicitly tasked with vetting CMC packages before submission. 

Brian Scanlan, managing partner at Freedom Bioscience Partners

that closing the speed gap with Chinese developers will depend on exactly this kind of regulatory agility paired with faster process development inside CDMOs. “Validated AI/ML models in process development at CDMOs can materially reduce development timelines,” Scanlan said, arguing that streamlined regulatory pathways and technology adoption need to advance together.

Manufacturing and quality organizations that partner with a sponsor or that serve as a QRI’s CMC reviewer should expect the pilot to test whether earlier, more collaborative CMC vetting genuinely reduces clinical holds without loosening FDA’s underlying quality expectations; the agency has stressed that it retains full authority over hold decisions throughout the program.¹

The pilot’s emphasis on pre-IND review has also drawn a response from those focused on what happens once a trial is underway. Richard Graham, chairman and co-founder, TruTechnologies, a clinical trial execution platform, said the pilot is a useful test of whether the FDA is willing to apply to itself the same risk-based, proportionate standard it has long asked of sponsors, including under ICH E6(R3), but that the effort needs to reach beyond trial start-up.

“This pilot is an opportunity for the FDA to apply that same principle to its own requirements,” Graham said. “The pilot could help make clearer what is actually required at a given stage of development and reduce work that does not add a commensurate safety benefit. Its attention to IRB review and site activation is welcome. But that thinking cannot stop at trial start-up. Clinical trial execution remains a major source of cost and delay, and proportionality has to extend into trial conduct more broadly.”

Graham added “That will require both a carrot and a stick: a clearer regulatory roadmap for what constitutes meaningful risk and the oversight those risks warrant, along with real accountability for identifying and managing the risks that truly matter. Nonbinding guidance alone is not going to move the needle. Technology exists to support that model. Sponsors can see whether protocol-required work is happening, where exceptions occur, and where attention is needed while there is still time to act. That enables more targeted oversight instead of blanket, retrospective checking. If we want the US to become a more attractive place to run clinical trials, faster regulatory review is only part of the answer. FDA has to create an environment where sponsors can confidently operate in the risk-based, proportionate way it has been asking them to for years.”

Because the initial cohort is capped at 8 to 10 pairs, and the FDA plans to use pilot results to inform a possible future QRI accreditation model, sponsors and CDMOs weighing an application have a narrow window to help shape whether rolling pre-IND CMC review becomes a standard part of FIH development going forward.¹

1. US Food and Drug Administration. FDA Launches Expedited IND Pilot, Begins Accepting Applications. Press Release. Sept. 15, 2026. 

2. US Food and Drug Administration. FDA Expedited Investigational New Drug (IND) Pilot Program. Accessed Sept. 16, 2026. 

3. CNBC. FDA Chief Warns U.S. Is Losing Ground to China in Early Drug Development, Calls for Faster Trial Approvals. Published Feb. 18, 2026. 

Articles

FDA has opened applications for its Expedited IND Pilot through Oct. 30, 2026, pairing sponsors with qualified research institutions to strengthen first-in-human submissions. The program enables rolling pre-IND review of IND components, aiming to surface deficiencies earlier and reduce clinical-hold risk once the 30-day review clock begins. Positioned within HHS’s Operation TrialBlazer, it reflects urgency to counter China’s faster early-development timelines. CMC is explicitly in scope, testing whether earlier, collaborative vetting accelerates starts without relaxing quality standards.

FDA opens applications for its Expedited IND Pilot, pairing sponsors with qualified research institutions to test rolling CMC review ahead of first-in-human trials.

FDA's IND Pilot Puts CMC Review First

At the 2026 PDA/FDA Joint Regulatory Conference, B. Timothy Bassler, SVP of Quality Operations at PCI Pharma Services, presented "Driving Quality Culture to Manage Risk," using a real-world FDA inspection case study to illustrate how quality culture and quality management systems (QMS) work together to manage compliance risk.

 His framing message: complexity is the enemy of compliance.

Bassler opened by outlining what active, visible quality oversight looks like in practice: routine presence on the shop floor, timely review of microbiological signals, willingness to challenge weak investigations, oversight of aseptic behaviors, and escalation of unresolved risks. He emphasized that a quality unit needs sufficient authority, technical expertise, staffing, and independence to act, including the authority to halt operations when sterility assurance may be compromised. That expectation tracks with how Siegfried Schmitt has 

: it's the function that implements and owns the pharmaceutical quality system underpinning a site's compliance, separate from, and independent of, the operations it oversees.

The presentation's centerpiece was a case study drawn from PCI's May 2025 acquisition of Ajinomoto Althea's San Diego-based sterile fill-finish site.

 Shortly after the acquisition closed, an FDA inspection at the site concluded on May 9, 2025, resulting in a Form 483 with multiple observations and triggering a formal remediation response.

 Bassler walked through the subsequent timeline: a July 2025 leadership meeting between FDA's Office of Manufacturing Quality and senior PCI leadership, voluntary commitments to remediate additional observations identified during that dialogue, the agency's issuance of Voluntary Action Indicated status, and ongoing status updates to FDA through completion of all committed actions. The remediation path Bassler described mirrors the pitfalls discussed during 

, when agency officials warned that companies too often propose quick fixes without thorough root-cause analysis instead of the comprehensive remediation that FDA expects.

A key theme of Bassler’s presentation was FDA's expectation that companies bring in third-party expertise to strengthen quality unit oversight, spanning audits of aseptic processing design, independent assessment of quality unit functions, and review of investigation systems for deviations, discrepancies, complaints, and out-of-specification results. Warning letters citing 

 after repeated aseptic-line failures have shown how quickly weak investigation systems escalate into agency action.

To operationalize FDA's feedback, PCI conducted a comprehensive site quality unit assessment combining document review, personnel interviews, a "voice of customer" survey, walkthroughs, mock disposition exercises, and trend analysis of quality metrics. That assessment produced seven defined actions across three workstreams: governance and authority, operational control, and capability-building for quality unit personnel.

Bassler detailed several resulting system improvements: an integrated batch disposition model embedding quality checkpoints throughout the batch lifecycle, an aseptic core monitor program providing real-time observation and peer-to-peer feedback on aseptic operations, a risk-tiered aseptic qualification framework, and a formal quality authority and decision-making matrix assigning investigation decision rights by role and risk.

Bassler closed with three takeaways: conduct a detailed current-state assessment before acting, set explicit expectations to drive process improvement, and track data continuously to sustain gains. The presentation credited PCI's San Diego Development and Manufacturing team, PricewaterhouseCoopers Advisory Services, and Network Partners Group for their roles in the remediation effort.



Parenteral Drug Association. PDA/FDA Join Regulatory Confernce 2026 Agenda. Available at 

https://www.pda.org/global-event-calendar/event-detail/pda-fda-joint-regulatory-conference-2026#agenda

PCI Pharma Services to Acquire US-based Aseptic Fill-Finish CDMO Ajinomoto Althea

Redica Systems. PCI Sand Diego, Inc. – Form 483, 2025-05-09. Available at 

https://www.redica.com/document-store/documents/view/100221964/pci-san-diego-inc-form-483-2025-05-09

Quality culture and an empowered, independent quality unit were positioned as the primary defenses against compliance risk in complex sterile manufacturing. Visible oversight means routine shop-floor engagement, rapid interpretation of microbiological signals, challenging weak investigations, monitoring aseptic behaviors, and escalating unresolved risks, including stopping operations. Following a May 2025 Form 483 at a newly acquired San Diego fill-finish site, leadership engaged FDA’s Office of Manufacturing Quality, made voluntary remediation commitments, achieved VAI status, and provided ongoing updates. Third-party expertise, structured assessments, and data-driven controls underpinned remediation.

At PDA/FDA Joint Regulatory Conference 2026, PCI Pharma's Bassler details how a QMS overhaul turned an FDA 483 into stronger quality culture.

Driving Quality Culture to Manage Risk: Lessons from an FDA Inspection Response

Based on Manufacturing Intelligence Episode 6, Chris Cole and Richard Jaenisch discuss the risks of using AI in Pharma Development and Manufacturing.

In episode 6 of the video series, Manufacturing Intelligence

 Chris Cole, spoke with Richard Jaenisch, senior director of education, outreach, and digital experience at Open Biopharma, about where artificial intelligence (AI) drug discovery actually stands, how the Pope's recent commentary on AI ethics compares to the embryonic stem cell debate, and where AI agents are, and aren’t, worth the security risk in manufacturing.

PharmTech: Where Is the Line Today of What AI Can Do in Discovery?

 There are really 2 approaches under the same umbrella. One mines existing data to identify gaps scientists might have missed, then applies that insight toward a new formulation. Scientists stay in control here, they can look at the output and say, ‘This makes sense,’ because they understand the underlying pathways. The other is closer to generating massive volumes of candidates and hoping something useful emerges, which I'd describe as chimpanzees writing Shakespeare. Give it enough attempts and you'll eventually get something, but I haven't seen that method produce a real success yet. The pathway-driven approach is where I expect real movement, since scientists can still translate and validate the results.

Are There Specific Capabilities Being Exaggerated?

It's not straightforward hype, and it's not fully delivering either. Even the newest generation of tools, like the latest AlphaFold, hallucinates more than its predecessor, and hallucination is inherently not useful, even though generating novel possibilities is part of the point of generative tools. Some of the gap comes down to how people use them. I've seen strong results from tools that weren't designed for the use case, simply because someone found a clever workaround. Benchmarks and real lab use don't always line up.

You've Compared the Pope’s Take on the AI Debate to the Stem Cell Controversy. Where Are You Seeing Those Parallels?

Both debates carry a strong moral charge that isn't really about the underlying science. With embryonic stem cells, the quandary centers on the source of the cells and where life begins. AI has no such component. It's a machine, not a form of life. What carries over is the pattern: a prominent religious voice shapes conversations well beyond people who share that faith. If colleagues in your circle are Catholic, AI ethics becomes a topic of conversation whether you are or not, and that attention compounds with other concerns already circulating, like the environmental footprint of data centers.

Where Are AI Agents Saving Time Today, And Where Is the Risk Not Worth It Yet?

Most agents in active use are doing simple work, scraping and compiling information, or handling short, sequential tasks. The majority deployed right now are coding agents, and that connects to manufacturing more than people expect, since every piece of equipment runs on software employees want to customize. The risk shows up with people who build bolt-ons without really understanding the code underneath, especially once an agent is given broad permissions and starts operating outside its intended scope.

Is There a Company Spending Enough Money to Make That Risk Worthwhile?

I'm not convinced yet. Announcing a partnership or a large infrastructure figure is easy, a $30 billion or $200 billion number tells me a company is buying capacity, but not that it's solving the training and talent problem underneath. I'll believe someone is serious when I see real money going into people doing the work, with the hours to match, not just infrastructure. Until that shows up in quarterly and investor reports, I'd call it mostly announcements rather than execution.

What Does a Pharma Company Doing This Right Look Like in Two Years?

We'll know more in 2 years, honestly. Companies are testing different strategies, and part of what separates them is whether the expertise they're hiring or training matches how they intend to use AI. My suspicion is that team-oriented implementations will outperform ones built around individual empowerment. When AI use is a shared, community resource rather than an individual one, tacit knowledge doesn't walk out the door when one person leaves, the rest of the team keeps it moving. Organizations that build AI adoption around teams, not just individuals, are the ones I expect real results from.

AI-enabled discovery is bifurcating into pathway-grounded, data-mining approaches that preserve scientific interpretability and high-throughput generative candidate churn with limited demonstrated success. Tool performance remains uneven; newer models can hallucinate, and clever user workflows often outperform benchmark expectations. AI-ethics discourse mirrors embryonic stem cell debates in moral salience and amplification by religious authorities, compounded by sustainability concerns. In manufacturing, agents mostly automate narrow tasks and code customization, but over-permissioned bolt-ons raise security risk. Capital outlays signal capacity, not talent; durable gains likely require team-centered adoption.

Richard Jaenisch, Open Biopharma breaks down AI's real limits in drug discovery, agent security risks, and what workforce readiness looks like by 2028.

Q&A: AI Coding Tools Carry the Real Risk

In the second half of a two-part deep dive on major pharma-AI partnerships with 

’s Chris Cole, Richard Jaenisch, senior director of education, outreach, and digital experience at Open Biopharma, argues that generative AI's promise in drug discovery is real but narrower than the hype suggests. Jaenisch draws a sharp distinction between AI tools that mine existing pharmaceutical datasets to surface overlooked structures — where scientists stay in control and can validate the science — and generative approaches he likens to "chimpanzees writing Shakespeare," where sheer volume substitutes for scientific rigor. The former is producing real pipeline candidates, including one drug now in Phase III; the latter, he says, hasn't yet demonstrated meaningful success. (

Watch Part One on why Roche's raw GPU infrastructure buildout and Sanofi's regional talent hub represent fundamentally different bets on where AI adoption actually stalls in pharmaceutical manufacturing.

In this episode, we discuss AI drug discovery hype, agent security risks, and why pharma's AI workforce bets remain unproven.

Manufacturing Intelligence: Sizing Up Pharma's Biggest AI Deals, Part Two 

2:17 – Where the line is today between what generative AI can do in discovery and formulation and what people imagine it does

9:38 – What’s feeding this disconnect, be it exaggerated capabilities or pure hype

11:16 – Similarities between the ethical conversations around AI and those around embryonic stem cell use

20:17 – Whether or not the Pope voicing opinions on AI has any sway, operationally, for pharma companies

22:09 – Where AI agents actually save time today in pharma where the security risk is just not worth it

31:23 – If there’s a pharma company or situation currently “hamfisting” enough money to make the security risk worth it

37:05 – With workforce repositioning as the metric that determines whether any of these investments pay off, what a pharma company doing this correctly looks like in two years

Jaenisch also draws an unexpected parallel between the current AI ethics debate and the embryonic stem cell controversy, noting that Pope Leo's recent commentary on AI has forced pharma leaders, Catholic or not, into conversations about morality and technology they weren't previously having. He connects this cultural moment to concerns about "AI psychosis" and over-reliance on chatbots as substitutes for human judgment and community.

On agentic AI specifically, Jaenisch is blunt about the security exposure created by ungoverned deployment. Most agents in active use today are simple scraping or reporting tools, he says, but coding agents that are increasingly used to build unauthorized software bolt-ons represent a growing risk, since the same capabilities that help non-experts build tools also help bad actors find exploits in them. He points to real-world incidents of agents breaking sandbox environments and warns that context-window limitations can cause agents to quietly drift from their original instructions over time.

Pressed on whether any pharma company has invested enough to meaningfully move the needle on AI-driven workforce transformation, Jaenisch is skeptical of headline infrastructure spending and argues the real signal will be sustained, team-based investment in people, not individual empowerment. "If everyone's stirring the bucket, the bucket keeps moving," he says, describing how team-oriented AI adoption preserves institutional knowledge even as individuals leave. He predicts the industry will have clearer answers on ROI and workforce strategy within two years, once companies begin reporting real financial results rather than partnership announcements.

The FDA on Aug. 5, 2026, approved Moderna's mFlusiva (mRNA-1010) for adults 50 and older, granting approval for two age groups (adults 50 to 65 and adults 65 and up) on the condition that Moderna run an additional clinical trial in the older group.

 The company said the vaccine will be available for this fall's flu season.

Moderna's mRNA shot was nearly 27% more effective against seasonal flu than a standard flu shot, according to clinical trial data from a study enrolling more than 40,000 adults aged 50 and older.

 Reactogenicity ran higher than with standard-dose comparators, including higher injection-site pain and systemic symptoms, though events were predominantly mild-to-moderate and resolved within one to two days.

 For adults aged 65 and older, immunogenicity data comparing mFlusiva against Fluzone High-Dose showed superior antibody responses at day 29 and at six months, the data underpinning that group's accelerated approval.

The review was not straightforward. The FDA initially refused this year to review Moderna's application before it reversed course a week later, with agency officials saying they wanted more data because Moderna had compared its vaccine with a standard flu shot in adults aged 65 and older whereas the CDC recommends a high-dose vaccine for that age group.

 An amended, age-stratified application followed, and the FDA's Vaccines and Related Biological Products Advisory Committee voted 9-0 that benefits outweighed risks ahead of the final decision.

That regulatory maneuvering echoes what our 

 series has tracked this year in tariff- and funding-driven shifts affecting mRNA programs, and it lands against a specific policy backdrop: HHS canceled roughly $500 million across 22 BARDA-funded mRNA vaccine projects in 2025, a decision that also forced a stop-work order on 

. The approval is likely to face pushback from allies of the health secretary, underscoring that mFlusiva's clearance came despite, not because of, the current federal posture toward the platform.

What Does the Approval Mean for Manufacturing?

mFlusiva extends Moderna's mRNA-lipid nanoparticle platform beyond COVID-19 into a second major respiratory indication, following the 2

That trajectory has been built on real infrastructure: Moderna's 

, all now underpinning a third product line rather than a single pandemic response.

The core manufacturing argument covered in our look at 

engineering challenges in commercial-scale mRNA production

—shared production infrastructure and rapid reformulation to match circulating strains—gets a real-world test as Moderna scales mFlusiva for fall 2026 distribution.

1. Lovelace, B. "FDA approves 1st mRNA flu shot, from Moderna." NBC News, Aug. 5, 2026. 

2. "FDA approves first mRNA-based flu vaccine." CNN, Aug. 5, 2026. 

3. "FDA Advisory Committee Supports Moderna mRNA Flu Vaccine Review." AJMC, 2026. 

4. "FDA Advisory Panel Votes 9-0 in Favor of Moderna's mRNA Flu Vaccine." BioPharm International, 2026. 

5. "A new type of flu shot is coming after FDA approves mRNA-based vaccine." Washington Post, Aug. 5, 2026. 

FDA approval of Moderna’s mFlusiva (mRNA-1010) introduces the first mRNA-based influenza vaccine for adults ≥50 years, with age-stratified authorization and a postmarketing trial requirement in those ≥65. In a >40,000-participant study, relative vaccine effectiveness was ~27% higher than a standard comparator, with increased but mostly mild-to-moderate, short-lived reactogenicity. For ≥65, accelerated approval relied on superior immunogenicity versus high-dose influenza vaccine at day 29 and six months. The decision arrives amid mixed federal signals toward mRNA programs.

FDA approves Moderna's mFlusiva, the first mRNA flu vaccine, testing the platform's manufacturing case amid federal funding cuts.