MIT study: 95% GenAI projects fail to show returns
A new MIT study, titled 'The GenAI Divide: State of AI in Business 2025,' which surveyed 300 AI deployments and spoke to 350 employees, found that 95% of AI projects showed no returns on investment for the companies.
The failure was not because AI models did not work efficiently, but because they were harder to adapt with pre-existing workflows in a company.
According to a Fortune report, citing the study's lead researcher, the 5% AI deployments that were successful were focused on solving a single problem and had worked on execution.
Companies are also dealing with a 'learning gap' in their workforce although company executives blamed AI model's performance.
The study also showed that more than half of GenAI budgets went to sales and marketing but the highest returns were seen in automating back-office processes like cutting external agency costs, and streamlining roles.
The report also noted that companies that purchasing AI tools from specialised vendors worked around 67% of the time while internal tools only succeeded one-third of the time.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
2 hours ago
- Time of India
Worried that AI will take away your job soon? Here is the good news from MIT
Selective displacement, not sweeping cuts Skills, not jobs, are the new battleground iStock Rather than eliminating large swaths of jobs, AI adoption is changing what employers look for in new hires. Debunking the biggest myths A slow shift, not a sudden storm Adapt, don't panic For years, headlines have warned of a looming wave of job losses driven by artificial intelligence. From call centers to classrooms, the fear has been that machines will replace humans faster than economies can adapt. But a new study from MIT suggests the reality is far less catastrophic — at least for report, titled The GenAI Divide: State of AI in Business 2025 , finds that while generative AI is reshaping certain business functions, it is not triggering mass layoffs. Instead, its impact on jobs is narrow, gradual, and researchers found that workforce reductions are limited to roles that were already considered vulnerable before AI entered the picture. Customer support, administrative processing, and standardized development tasks — many of which had been outsourced in the past — account for most of the reductions, typically in the range of 5 to 20 energy, and advanced industries, by contrast, reported no significant reductions in hiring. Executives in these sectors told MIT they do not anticipate cutting back on critical roles such as physicians or clinical staff over the next five years. The only sectors bracing for noticeable workforce shifts are technology and media, where more than 80 percent of executives expect reduced hiring volumes within the next two than eliminating large swaths of jobs, AI adoption is changing what employers look for in new hires. Companies across industries are increasingly prioritizing candidates with AI literacy.'Our hiring strategy prioritizes candidates who demonstrate AI tool proficiency. Recent graduates often exceed experienced professionals in this capability,' a VP of Operations at a mid-market manufacturing firm told MIT reflects a broader recognition that AI proficiency is no longer an optional skill but a competitive advantage in modern MIT study also dismantles some of the most common fears surrounding AI in the workplace. Among the five myths it highlights:Many assume AI will replace most jobs within a few years, but the evidence shows layoffs are limited and concentrated only in certain industries. Another belief is that generative AI is already transforming business. While adoption is widespread, real transformation is rare, with only 5 percent of enterprises integrating AI at contrary to the notion that large firms are slow adopters, the report shows 90 percent have actively explored AI solutions. The real problem is not model quality or regulations but the fact that most AI tools do not learn or integrate seamlessly into workflows. Finally, the idea that the best companies build their own tools is challenged by data showing internal builds fail at twice the rate of external to MIT's Project Iceberg analysis, only 2.27 percent of U.S. labor value is currently exposed to automation. While the long-term potential for AI to reshape as many as 39 million jobs exists, the researchers emphasize that such transformations will take time. Until AI systems develop persistent memory and autonomous learning, their influence will be limited to cost optimization rather than wholesale restructuring of contrasts sharply with Wall Street expectations. Investors have poured more than $44 billion into AI startups in just the first half of 2025, betting on unprecedented productivity gains. Goldman Sachs estimates total AI investment could hit nearly $200 billion this year. Yet as MIT and Fortune note, 95 percent of AI business experiments are failing to deliver measurable MIT report makes clear that the AI revolution is more of a slow burn than an overnight upheaval. For workers worried about their jobs vanishing tomorrow, this is good news. The bigger shift lies not in disappearing roles, but in evolving other words, AI is unlikely to replace humans en masse in the immediate future. Instead, it is nudging workplaces toward a new reality where the ability to collaborate with machines may be just as important as traditional if you are worried that AI will take away your job, MIT's research suggests a more practical response: learn the tools, embrace the changes, and position yourself on the right side of the AI divide.
Time of India
2 hours ago
- Time of India
Looking for job switch? MIT study reveasls what employers scan first on your resume
The real job impact: selective, not universal What employers want most in candidates iStock The biggest shift is not in how many people are hired, but in who gets hired. A slow shift, not an overnight takeover The bottom line for job seekers You Might Also Like: Worried that AI will take away your job soon? Here is the good news from MIT If you are polishing your resume for the next big career move, here is some good news: artificial intelligence may not be taking away jobs as quickly as many fear. A new MIT report, The GenAI Divide: State of AI in Business 2025, reveals that while generative AI is changing hiring patterns, it is not triggering the kind of sweeping layoffs that were once predicted. Instead, employers are evolving their expectations, and one skill is emerging as the make-or-break factor for candidates — AI MIT researchers found that workforce reductions due to generative AI are concentrated in narrowly defined functions such as customer support, administrative processing, and standardized development tasks. These areas, which were already prone to outsourcing, saw headcount reductions between 5 and 20 percent among advanced AI of tech-heavy industries, however, executives expressed little concern about large-scale job cuts. In healthcare, for example, leaders reported no expectation of reducing hiring for physicians or clinical staff. Similarly, sectors like energy and advanced manufacturing remain largely insulated from AI-driven disruption. By contrast, more than 80 percent of executives in technology and media anticipate cutting back on hiring volumes within the next two biggest shift is not in how many people are hired, but in who gets hired. Across industries, the MIT study highlights a growing consensus: candidates with proficiency in AI tools are being prioritized over those without.'Our hiring strategy prioritizes candidates who demonstrate AI tool proficiency. Recent graduates often exceed experienced professionals in this capability,' a Vice President of Operations at a mid-market manufacturing firm told MIT underscores a broader trend: employers are not simply seeking technical specialists, but rather employees at every level who are comfortable working alongside AI systems. In practice, this means listing your experience with AI-powered platforms, from automation tools to generative AI assistants, could carry as much weight as traditional 'Project Iceberg' analysis shows that while AI has the potential to automate as much as $2.3 trillion in U.S. labor value, the actual automation happening right now is just 2.27 percent of labor output. The researchers conclude that workforce transformation will unfold gradually, tied to AI's ability to develop memory, contextual learning, and autonomy — features that most systems still short, the robots are not replacing you tomorrow. Instead, companies are focusing on external cost savings and incremental efficiency rather than mass professionals considering a switch, the takeaway is clear: AI skills are quickly becoming the new literacy of the workplace. Knowing how to navigate tools like ChatGPT, Copilot, or specialized AI platforms can put candidates ahead of the curve. At the same time, the human touch in problem-solving, creativity, and contextual judgment remains irreplaceable — at least for while the AI hype has investors pouring billions into the technology, job seekers should not panic. Instead, they should see it as a nudge to upskill. As the MIT report makes clear, the most attractive resumes in 2025 are not the ones replacing humans with AI, but the ones showing how humans can work smarter with it.
Hindustan Times
2 hours ago
- Hindustan Times
How Generative AI can improve patient care
Artificial Intelligence (AI) tools are transforming industries, such as finance and banking, retail, and telecommunications, by boosting efficiency and quality. The potential of AI lies in its adaptability and customisation, which have vast benefits for health care management and clinical practice. Representational Image(Pixabay) While AI technologies in imaging, radiology, and precision medicine hold significant promise for transforming health care, their integration into the Indian health care system remains limited due to low levels of hospital digitalisation and a lack of widespread AI literacy. However, with the undeniable benefits offered by AI tools for patients and clinicians alike, adoption is gradually gaining momentum in the Indian health care landscape. The 2023 guidelines on ethical and responsible use of AI by Indian Council of Medical Research (ICMR) were a welcome step to pave the way for safe, human-centric AI integration. Generative AI demonstrates significant potential in health care by optimising medical workflows, supporting evidence-based clinical decision-making, and improving patient outcomes. The deployment of specialised, use-case-specific AI tools enables health care systems to enhance operational efficiency while advancing the quality of care. Hospitals are intricate ecosystems that demand seamless coordination among administrative staff, clinicians, surgeons, radiologists, nurses, and support services. Medical management teams are tasked with ensuring quality and safety, overseeing the synergy between different departments, and facilitating optimal resource utilisation. Adoption of electronic health records (EHR) and AI-optimized hospital information management systems can enable GenAI tools to create impact through strategic support. For instance, adoption of GAI-powered conversational chatbots is suitable and effective in various clinical settings. These chatbots can conduct pre-consultation interviews with patients, which saves time for both clinicians and patients. By capturing medical histories and relevant details in advance, they enable physicians to focus more on meaningful patient interaction during consultations. GAI-based clinical assistants also contribute significantly to tasks such as medicine reconciliation and auto-reporting for investigations, which can be vetted and approved by the treating physician. By automating tasks like test ordering, communication between departments becomes faster and less burdensome for doctors. Beyond administration, GenAI supports evidence-based medicine, transforming clinical decision support by enabling more dynamic, context-aware, and patient-centred care. GAI-powered Clinical Decision Support Systems (CDSS) works as a highly specialised search engine that draws on a database of only peer-reviewed, verified, and high-quality sources, conforming to strict medical guidelines and standards. This prioritises the Human-in-the-Loop (HITL) approach, where human expertise and knowledge to make critical decisions are supported by GAI capabilities.. Traditionally, clinicians relied on extensive textbooks or unverified online sources for medical research. Today, GAI-powered Clinical Decision Support Systems (CDSS) offer instant access to the latest, evidence-based clinical data, enabling personalised treatment planning with precision. While CDSS does not replace medical expertise, it enhances it--empowering doctors and nurses to make faster, better-informed decisions backed by verified sources and transparent citations, ultimately improving patient care. There's much that CDSS and GenAI can do, but their limits are clearly defined. The role of AI in health care will always be to support and enhance. Values like accountability, empathy, sensitivity, and compassion are the domain of human beings. AI tools for health care must be ethically designed and sensitively implemented with a non-negotiable focus on safety and quality, allowing health care professionals to do the best possible work they can. As GAI uptake increases and becomes more accessible, such tools have the potential to greatly benefit medical students, doctors in remote and underserved areas, ASHA workers, and community physicians – making health care more accessible, accurate, and patient-centric across India. This article is authored by Nimisha Goswami, director, content strategy, Elsevier.
