Technology & innovation
The majority of pharmaceutical, biotech and bioengineering firms in the U.S., Canada and Europe are pursuing agentic AI initiatives, according to a new survey published by the MIT Technology Review.
Agentic AI – the technology that is capable of executing “multi-step tasks under policy constraints with human sign-offs” – is still relatively new, but Gartner recently predicted “at least 15% of day-to-day work decisions will be made autonomously by AI agents by 2028.”
Highlights from the survey of 250 executives from pharma and biotech firms across the U.S., Canada, France, Germany, Italy, Spain, UK, Brazil, and Switzerland:
How many pharma/biotech firms are pursuing agentic AI, and how far along are firms in the implementation process?
73% of surveyed firms say they are "planning, piloting, or deploying agentic AI initiatives in commercial and medical affairs operations."
That said, the majority of firms surveyed by MIT Technology Review are "pursuing agentic AI initiatives," with 7% of firms already deploying agentic AI in production.
22% are in the pilot stage, with agentic AI being tested across "multiple commercial/ medical functions," while 16% of firms are "running limited proof-of-concept initiatives in one or two areas."
29% of firms are in the "planning/evaluation phase" for agentic systems they have yet to implement.
More than a quarter (27%) are "currently abstaining from agentic AI," including 14% of firms "using traditional AI/automation tools but not agentic systems" and 13% of firms that are “not currently pursuing agentic AI initiatives.”
What are the challenges around implementing agentic AI initiatives?
60% of surveyed leaders cited the "complexity of designing, orchestrating, and managing advanced agentic workflows" as a "critical or major barrier" in going from pilot to a wider rollout.
55% of leaders cited "navigating regulatory compliance and validation requirements" as a significant barrier, while 42% of respondents cited "technology infrastructure limitations and system operability."
Only 21% of leaders cited “organizational resistance to AI-human workflows a critical or major barrier.”
Read more via MIT Technology Review
According to a new report by SHRM, a significant percentage of U.S. jobs are "at least 50% automated" and therefore are at risk of displacement.
Highlights from the SHRM 2025 Automation/AI Survey, including responses from more than 20,000 U.S. workers:
15% of U.S. jobs hit a "high automation" threshold. SHRM "identified jobs as possessing a high automation level if at least 50% of the tasks in the job are automated." By that definition, SHRM estimates that "15.1% of U.S. employment is at least 50% automated." The share that meets the 50% automation threshold "varies significantly by major occupational group, from a low of 7.3% (education and library occupations) to a high of 32% (computer and mathematical occupations)."
The share of jobs that are at risk of displacement “varies significantly by industry.” Just 10.7% of roles in educational services meet the 50% automation threshold, compared to 21.7% of information roles. Sectors that "heavily rely on computer and/or software systems that are becoming increasingly sophisticated and autonomous" tend to have a higher share of "employment that is at least 50% automated," compared with sectors that are more heavily reliant on "direct interpersonal engagement."
7.8% of U.S. employment is "at least 50% done using generative AI." SHRM notes that "there is significant variation in this share across occupational groups, from a low of 4.3% (health care practitioners) to a high of 19.2% (computer and mathematical occupations)."
Nontechnical barriers to automation displacement still exist: While 15% of U.S. jobs are identified as meeting a high automation level, SHRM notes that there are "nontechnical barriers to automation displacement" that may be protecting jobs from displacement. Such barriers can include client preferences and legal and regulatory requirements. (Still, 6% of U.S. employment is “at least 50% automated and has no definitive nontechnical barriers to automation displacement.”)
Read more via SHRM, HR Dive
AI is now impacting the way both job seekers and hiring managers approach the process. That much is clear.
Organizations are using AI to "screen thousands of candidates." Job seekers are leaning on AI to tailor their resume to a particular role or to write cover letters. But when it comes to the interview process, is it okay for candidates to lean on AI to help them with submissions for assignments designed to gauge their skills?
Stacy Martinet, Adobe's Chief Communications Officer, says that she "favors applicants who apply the tools directly to their work." Martinet says that "during interviews or test projects, she values submissions enhanced by AI, describing it as a mark of adaptability and innovation."
Until recently, Anthropic "barred candidates from using any AI assistants when answering application questions, citing a need to evaluate unassisted communication skills." (That policy changed in July, according to HR Grapevine.)
Goldman Sachs prohibits candidates from using generative AI during interviews.
Amazon "warned recruiters that any candidates who are caught using the tech should be disqualified," according to Business Insider.
Read more via HR Grapevine, Futurism, Business Insider
The U.S. Air Force is running “Top Gun–style” tests where pilots train alongside AI-piloted XQ-58 Valkyrie drones. The AI system can make split-second decisions and is now learning defensive maneuvers to support human pilots in future missions. These tests at Eglin Air Force Base aim to refine human-machine teaming, with Valkyries flying alongside F-35s, F-22s, and F-15EXs. Officials say the program could pave the way for broader deployment of semi-autonomous fighter capabilities across the U.S. Air Force. (InterestingEngineering.com)
The "agentic AI revolution" is "going to be magic," according to Salesforce CEO Marc Benioff. Speaking at the recent Dreamforce conference, Benioff "outlined his vision for how autonomous, "agentic" AI systems will reshape business operations, from marketing to manufacturing." (Yahoo Finance)
