Commonly used generative AI models, such as ChatGPT and DeepSeek R1, are highly vulnerable to repeating and elaborating on medical misinformation, according to new research.
Mount Sinai researchers published a study this month revealing that when fictional medical terms were inserted into patient scenarios, large language models accepted them without question — and went on to generate detailed explanations for entirely fabricated conditions and treatments.
Even a single made-up term can derail a conversation with an AI chatbot, said Dr. Eyal Klang, one of the study’s authors and Mount Sinai’s chief of generative AI. He and the rest of the research team found that introducing just one false medical term, such as a fake disease or symptom, was enough to prompt a chatbot to hallucinate and produce authoritative-sounding — yet wholly inaccurate — responses
Dr. Klang and his team conducted two rounds of testing. In the first, chatbots were simply fed the patients scenarios, and in the second, the researchers added a one-line cautionary note to the prompt, reminding the AI model that not all the information provided may be inaccurate.
Adding this prompt decreased hallucinations by about half, Dr. Klang said.
The research team tested six large language models, all of which are “extremely popular,” he stated. For example, ChatGPT receives about 2.5 billion prompts per day from its users. People are also becoming increasingly exposed to large language models whether they seek them out or not — such as when a simple Google search delivers a Gemini-generated summary, Dr. Klang noted.
But the fact that popular chatbots can sometimes spread health misinformation doesn’t mean healthcare should abandon or scale back generative AI, he remarked.
Generative AI use is becoming more and more common in healthcare settings for good reason — because of how well these tools can speed up clinicians’ manual work during an ongoing burnout crisis, Dr. Klang pointed out.
“[Large language models] basically emulate our work in front of a computer. If you have a patient report and you want a summary of that, they’re very good. They’re very good at administrative work and can have very good reasoning capacity, so they can come up with things like medical suggestions. And you will see it more and more,” he stated.
It’s clear that novel forms of AI will become even more present in healthcare in the coming years, Dr. Klang added. AI startups are dominating the digital health investment market, companies like Abridge and Ambience Healthcare are surpassing unicorn status, and the White House recently issued an action plan to advance AI’s use in critical sectors like healthcare.
Some experts were surprised that the White House’s AI action plan didn’t have a greater emphasis on AI safety, given it’s a major priority within the AI research community.
For instance, responsible AI use is a frequently discussed topic at industry events, and organizations focused on AI safety in healthcare — such as the Coalition for Health AI and Digital Medicine Society — have attracted thousands of members. Also, companies like OpenAI and Anthropic have dedicated significant amounts of their computing resources to safety efforts.
Dr. Klang noted that the healthcare AI community is well aware about the risk of hallucinations, and it is still working to best mitigate harmful outputs.
Moving forward, he emphasized the need for better safeguards and continued human oversight to ensure safety.
Jobs aren’t disappearing. Through AI, talent acquisition is fast evolving from hand-crafted interviewing and recruiting to a data-driven model that ensures the right talent is hired at the right time, for the right role with unmatched accuracy
Traditional recruiting isn’t working: in 2024, only 17% of applicants received interviews and 60% abandoned slow application processes
AI drives 2–3x faster hiring, stronger candidate quality, sharper targeting—and 95% candidate satisfaction at Foundever, from 200,000+ applicants in just six months
OAKLAND, Calif., Sept. 16, 2025 /PRNewswire/ — The Josh Bersin Company, the world’s most trusted HR advisory firm, today released new research showing that jobs aren’t disappearing—they’re being matched with greater intelligence. The research, produced in collaboration with AMS, reveals major advances in talent acquisition (TA) driven by AI-enabled technology, which are yielding 2–3x faster time to hire, stronger candidate-role matches, and unprecedented precision in sourcing.
The Josh Bersin Company (PRNewsfoto/The Josh Bersin Company)
The global market for recruiting, hiring, and staffing is over $850 billion and is growing at 13% per year, despite the economic slowdown, though signs of strain are evident. This means TA leaders are turning to AI to adapt, as AI transforms jobs, creates the need for new roles, new skills, and AI expertise.
According to the research and advisory firm, even without AI disruption, over 20% of employees consider changing jobs each year, driving demand for a new wave of high-precision, AI-powered tools for assessment, interviewing, selection, and hiring. Companies joining this AI revolution are hiring 200-300% faster, with greater accuracy and efficiency than their peers, despite the job market slowdown.
According to the report, The Talent Acquisition Revolution: How AI is Transforming Recruiting, the TA automation revolution is delivering benefits across the hiring ecosystem: job seekers experience faster recognition and better fit, while employers gain accurate, real-time, and highly scalable recruitment.
This is against a context of failure with current hiring. In 2024, less than one in four (17%) of applicants made it to the interview stage, and 60% of job seekers, due to too-slow hiring portals, abandoned the whole application process.
The research shows how organizations are already realizing benefits such as lower hiring costs, stronger internal mobility, and higher productivity. AI-empowered TA teams are also streamlining operations by shifting large portions of manual, admin-heavy work to specialized vendors.
Early successes are striking: after deploying conversational AI, a major U.S. resort operator increased scheduled interviews by 423% in 12 months while reducing candidate drop-off by 85%. A new AI TA process at Foundever hit 95% candidate satisfaction rating from over 200,000 applicants in just six months, while a leading global automotive company reported $2 million in savings in its first year using AI-powered interview scheduling.
The research highlights how AI is helping TA overcome frustrations like vague job descriptions, inconsistent interviews, and labor-intensive processes.
HR teams now use AI to automate tasks from profiling and sourcing to screening, scheduling, interviewing, and negotiating offers. Companies showcasing these successes include major international brands spanning the hospitality, food, healthcare, and technology sectors, such as Fontainebleau Las Vegas and Compass Group.
Some organizations are using AI-powered recruiting assistants to manage routine communication and negotiations. Recruiters set parameters for salary, benefits, and start dates, while the AI answers questions, updates offers, and proposes alternatives in real time. By automating tasks and personalizing the experience, AI shortens turnaround times and frees recruiters to focus on strategy.
The vendor market is transforming, with SAP acquiring SmartRecruiters and Workday acquiring Paradox to stay competitive. Innovations include AI agents enabling automation of the full hiring journey (Eightfold AI, Maki People) conversational AI for candidate engagement (Glider AI, Paradox, Radancy), AI-driven assessments (CodeSignal, HireVue, TaTiO), and platforms that benchmark roles against labor market data (Draup, Galileo, Lightcast, Reejig, SeekOut, TechWolf).
Report author and Josh Bersin Company Industry Analyst & Senior Research Director, Stella Ioannidou, says:
“For decades, TA has been viewed as a cost center, focused on using applicant tracking systems to manage incoming candidates and relying on recruiters to screen and interview. This process was slow, expensive, and delivered a poor experience for job seekers.
“Today, by leveraging AI-powered platforms and integrated data, TA teams can identify, attract, and engage the best talent in the market with unparalleled precision, often before competitors even know those candidates are available—resulting in a proactive, data-driven approach that enables organizations to respond quickly to changing business needs, seize new opportunities, and fuel growth from within.”
Chief Executive Officer at AMS, Gordon Stuart, says:
“This research paper captures the urgency and scale of the AI revolution which is transforming TA. It doesn’t just reflect what’s happening now, it helps us understand what’s next. Candidates want speed, clarity, and connection. Recruiters need tools that free them to focus on strategy and relationships. And businesses must rethink how talent acquisition fits into their broader growth agenda. AI is not just automating tasks, it’s redefining roles, workflows, and expectations across the board. The cumulative power of people, process, data and technology is heralding a new era for talent acquisition.”
Global industry analyst and Josh Bersin Company CEO, Josh Bersin, says:
“AI is expected to provide CHROs with a data-rich view of talent comparable to an integrated supply chain, enabling them to track and analyze every detail of each hire with the same precision a luxury Swiss watchmaker applies to every component and its origin.
“This transition transforms HR from handcrafted processes to precision, dynamic hiring—something once unattainable without AI.
“The implications are significant for all stakeholders, particularly CEOs. Organizations that lacked a precision hiring process have historically faced millions in lost revenue, high turnover, and misaligned talent—but those days are now coming to an end.”
This new research follows previous Josh Bersin Company findings demonstrating how AI is transforming another core HR function: Learning & Development.
The Josh Bersin Company’s Galileo Suite delivers both strategic guidance and hands-on tools through its AI Agent for HR, as well as hyper-personalized learning through the exclusive Galileo Learn certificate course, AI-First TA Transformation. The full report is available to download here.
About AMS AMS is a leading global provider of talent acquisition and consulting services, providing unrivalled experience, driven by technology and underpinned by innovation. We help our clients to attract, engage and retain the talent they need for business success.
We have three core areas of service: acquisition, advisory and digital, mainly delivered as an outsourced model, and spanning our clients’ permanent and contingent workforce, and internal mobility requirements.
Our dedicated teams of experts are deeply embedded with our global blue-chip clients, enhancing talent acquisition processes and driving projects which align with overall strategic objectives. This relationship-driven approach is supporting our clients to redefine how they hire and retain top talent. For more, go to www.weareams.com/
About The Josh Bersin Company
The most trusted human capital advisors in the world. More than a million HR and business leaders rely on us to help them overcome their greatest people challenges.
Thanks to our understanding of workplace issues, informed by the largest and most up-to-date data sets on workers and employees, we give leaders the confidence to make decisions in line with the latest thinking and evidence about work and the workplace. We’re great listeners, too. There’s no one like us, who understands this area so comprehensively and without bias.
Our offerings include the industry’s leading AI-powered HR expert assistant, Galileo®, fueled by 25 years of in-depth Josh Bersin Company research, case studies, benchmarks, and market information.
We help CHROs and CEOs be better at delivering their business goals. We do that by helping you to manage people better. We are enablers at our core. We provide strategic advice and counsel supported by in-depth research, thought leadership, and unrivaled professional development, community, and networking opportunities.
We empower our clients to run their businesses better. And we empower the market by identifying results-driven practices that make work better for every person on the planet.
Artificial intelligence (AI) is often criticised for ripping off artists, but the technology is now being used to combat fake copies of works by the Canadian artist Norval Morrisseau (1932-2007) that have flooded the market over the past two decades.
More than 6,000 pieces were produced and fraudulently sold as authentic works by the Ojibwe artist to collectors worldwide, with financial losses estimated to exceed C$100m ($72.5m). The trial of Jeffrey Cowan, the last of the suspected fraudsters in the tangled web of Morrisseau forgery rings, began this week. Two other men who pleaded guilty to participating in what the Ontario Provincial Police described as “the biggest art fraud in world history”, were each given a conditional sentence of two years less a day in August and September.
In a unique case of fighting fakes with fakes, the Montreal-based start-up Acrylic Robotics announced in July that, in partnership with the Norval Morrisseau Estate, its robots will reproduce five Morrisseau paintings and make the copies available for purchase.
Norval AI
“We have created our own artificial intelligence called Norval AI to help determine the probability of an authentic Norval Morrisseau painting,” Cory Dingle, the Morrisseau estate’s executive director, tells The Art Newspaper. “It has grown to do many other functions that will help with museums seeking provenance as well as law enforcement—such as catching the person who painted the fraudulent work.”
The idea of developing an AI authenticator began in 2023, when Dingle met Stephan Heblich, an economics professor at the University of Toronto, and Clément Gorin, an associate professor at the Université Paris 1 Panthéon-Sorbonne.
“Clément and I were just two art-loving economics professors who are using the latest deep learning and visual recognition techniques to analyse paintings,” Heblich says. “We thought we could help restore Norval’s legacy, so we reached out to Cory and created Norval AI.”
A year later, Dingle met Chloë Ryan, the founder and chief executive of Acrylic Robotics. The artist-turned-engineer had developed technology that allows robots to paint works in the style of individual artists.
Better fakes
“We needed better replicas to test our artificial intelligence programme—the existing fake paintings were so terrible,” Dingle says. “So we collaborated with Acrylic Robotics and helped train their robot to paint more realistic fake paintings.” He adds: “They want to produce very accurate reproductions and our Norval AI tells their robot where it is doing a bad job. They use our data to make the robot do a better job, which makes those better replicas train our Norval AI even better.”
Last year, the Norval Morrisseau Family Foundation helped Acrylic Robotics produce a very accurate replica of one of Morrisseau’s original paintings. This in turn was run through the Norval AI programme and the results were shared with Acrylic Robotics to help improve its replicas.
The resulting works produced by Acrylic Robotics include limited editions of five paintings, including Morrisseau’s In Honour of Native Motherhood (1990), which was inspired by the murdered and missing Indigenous women in Canada, and Punk Rockers (around 1991), in which Morrisseau fused traditional Anishinaabe iconography with contemporary idioms.
Marked as replicas
Prices for the Acrylic Robotics works range from C$3,240 to C$45,000. To avoid further fraud, several techniques have been applied to ensure the works are easily identifiable as replicas, including a mark on the back of the canvas.
According to an Acrylic Robotics spokesperson, the company hopes to investigate whether, with the support of Norval AI, its robots may be able to complete some of Morrisseau’s many unfinished or damaged works.
“What’s exciting here, even beyond the technology,” Ryan says, “is the opportunity to push the boundaries of what has been possible, make art history and reclaim a legacy.”
If you buy the big tech hype around AI, you’d be forgiven for assuming that the future will be plain sailing with agents carrying out drudge work while you focus on the nitty gritty of your job.
Jensen Huang, CEO at Nvidia, isn’t convinced on that, however. You’ll probably still be bogged down, overworked, and wondering exactly why billions of dollars have been poured into a technology that’s making your life a misery.
In a recent appearance on Fox Business, Huang responded to a question on robotics and AI productivity gains with a rather grim outlook: workers will actually be busier.
“I have to admit that I’m afraid to say that we are going to be busier in the future than now,” he told host Liz Claman. “And the reason for that is because a lot of different things that take a long time to do are now faster to do. I’m always waiting for work to get done because I’ve got more ideas.”
“The more productive we are, the more opportunity we get to pursue new ideas,” Huang continued.
Reading between the lines here, it seems the so-called efficiency gains afforded by AI will mean workers have more work dumped in their laps – onto the next task, no rest for the wicked, etc.
Huang’s comments run counter to the prevailing sentiment among big tech executives on exactly what AI will deliver for both enterprises and individual workers.
Sign up today and you will receive a free copy of our Future Focus 2025 report – the leading guidance on AI, cybersecurity and other IT challenges as per 700+ senior executives
Since the advent of the generative AI ‘boom’, the technology has been hailed as a silver bullet for sluggish performance, a means to supercharge worker productivity and maximize efficiency.
There are signs the technology is delivering on these fronts. Analysis from GitHub last year showed that software developers were saving so much time using AI tools that they were able to focus on other tasks, with some even taking the time to learn new programming languages.
Separate research from Microsoft also points toward positive productivity gains for individual workers. Naturally, productivity boosts are welcomed. However, Huang’s comments point to a more concerning reality with the deployment of AI in the enterprise.
Nobody is expecting the technology to pick up slack and enable workers to put their feet up. But if the technology is being used efficiently and unlocking gains, and we’re still busier, then the messaging we’ve had from big tech over the last two years has been false.
Expectations versus reality
We’ve all read the marketing copy and heard it regurgitated by tech leaders on podcasts and keynote stages – AI will allow us to focus on the “more rewarding” aspects of our jobs.
They’ve never fully explained what this entails, or how it will pan out in the workplace. To be quite honest, I don’t think they know what it means. Marketing probably made it up and they’ve stuck with it.
Moreover, AI has been touted as the perfect tool to help reduce manual toil and ease the burden on staff. If the technology delivers on this promise, then why will it be busier for workers?
Will we be busier spending time on those rewarding aspects of our jobs? I have to say, I’m doubtful. The reality is that workers will be pulled into other tasks and merely end up drowning in the same cumbersome workloads they’ve been dealing with since the pandemic.
To Huang’s credit, he does mention the prospect of a four-day week becoming a reality in this AI-powered future. That’s something I think most workers would agree is a positive.
“I fully expect GDP to grow. I expect productivity to increase. I actually expect us to have more things to do,” he said.
“I’m hopeful for that day too, so that we have four-day work weeks… so that we could spend more time on the weekends with family and get some reading done, and do some travelling. Nothing is better than that.”
In this sense, AI could act as an enabler in the natural evolutionary path of working practices, which as Huang points out we’ve already seen change drastically over the last few decades.
“We came from a world of seven-day work weeks,” Huang noted. “Now we’re in five-day work weeks. And every industrial revolution leads to some change in social behaviour. I expect the economy to be doing very well because of AI and automation.”
I’m still not buying it
Call me a cynic, but dreams of a more productive, AI-powered future where workers have four-day weeks and spend time with family sound too good to be true.
Dell Technologies, Amazon, and Google have all cracked down with strict RTO mandates, with Microsoft reported to be planning a similar move. If AI deployments successfully lighten employee workloads, do you think executives will be content?
Absolutely not. They’re hellbent on squeezing every possible dollar out of the workforce and equally hellbent on justifying the billions they’ve invested in the technology.
AI might help workers do their jobs more efficiently, but if anything that will likely justify further workforce reductions. If ten people have so much spare time, why not just cut them down to five?
Microsoft, Cisco, and Oracle have all announced workforce reductions in recent months despite posting strong financial results. If they’re willing to cut staff in these circumstances, they’re willing to do it any time.
