Time saved by AI offset by new work created, study suggests

“The adoption of these chatbots has been remarkably fast,” Humlum told The Register about the study. “Most workers in the exposed occupations have now adopted these chatbots… But then when we look at the economic outcomes, it really has not moved the needle.” AI start-ups such as Character.AI that offer chatbots as “companions” have faced criticism for allegedly not doing enough to protect users. Last year, a teenager killed himself after interacting with Character.AI’s chatbot. The teen’s family is suing the company for allegedly causing wrongful death, as well as for negligence and deceptive trade practices.

Shifting Model Preferences

• These systems might not be as flashy as gen AI, but classic artificial intelligence is a huge part of the technology we rely on every day.
• Leo CybSec unites a group of Cyber Security experts with 20+ years of collective expertise to help our clients realise and mitigate the cyber challenges and risks facing their business.
• Many people have interacted with chatbots in customer service or used virtual assistants like Siri, Alexa and Google Assistant — which are now on the cusp of becoming gen AI power tools.
• Finally, teams often underestimate the human expertise required, or they get stuck deciding whether to build in-house or buy from providers.
• For example, defend — the most popular framework — could involve customer service chatbots that reduce churn by improving response times.

The software can then come up with teams of AI agents, ready to be managed by human users, under the governance and guardrails organizations put in place to ensure they perform the right tasks in the right way. Within these parameters, users can test how the agents perform in real-world use cases, then set them to work. Many people have interacted with chatbots in customer service or used virtual assistants like Siri, Alexa and Google Assistant — which are now on the cusp of becoming gen AI power tools. All that, along with apps for ChatGPT, Claude and other new tools, is putting AI in your hands. Many users enjoy the convenience and creativity it offers, especially for things like writing help, image creation, homework support and productivity. “This new capability in our Automation Co-Pilot represents a new era in enterprise AI—one where people don’t just use automation, they collaborate with it,” said Adi Kuruganti, chief product officer at Automation Anywhere.

Nearly half of agentic AI projects will be killed by ’27 due to hype, costs, and risks

Cost and budget constraints, noted by 24%, also pose challenges, as LLM deployment demands substantial investment in cloud infrastructure and fine-tuning. A new report from Kong, titled “What’s Next for Generative AI in the Enterprise,” reveals that 72% of enterprises plan to increase their spending on large language models (LLMs) in 2025, with nearly 40% expecting investments to exceed $250,000 annually. Despite finding widespread and often employer-encouraged adoption of these tools, the study concluded that “AI chatbots have had no significant impact on earnings or recorded hours in any occupation” during the period studied. The confidence intervals in their statistical analysis ruled out average effects larger than 1 percent. With open-source models, the Indian government aims to make AI more understandable, adaptable, and accessible to the country’s local needs.

However, challenges such as security concerns and integration complexities loom large. Even so, the Danish study provides a valuable but limited snapshot that challenges some narratives of immediate, widespread labor market transformation from generative AI. Given the rapid pace of AI development, the longer-term economic impact of generative AI remains an uncertain and debated question that will likely be the subject of many future research papers. The issue, stemming from how the large language models are trained, has come into focus at a time when more and more people have adopted the chatbots not only at work as research assistants, but in their personal lives as therapists and social companions. Various generative AI tools and services have gradually permeated everyday life and the workplace, driving rapid market growth and attracting startup teams to enter the field.

Unclear ROI, weak governance, and vendor lock-in further complicate the picture. Enterprises are optimistic about AI’s transformative potential, with 82% of respondents believing GenAI will positively impact their careers and organisations. The report identifies productivity and innovation as key benefits, with 46% viewing AI as a tool to streamline tasks like coding, documentation, and API testing. As businesses race to harness AI’s potential, the report underscores a shift towards integrating LLMs into core operations, driven by the promise of enhanced productivity and innovation.

In customer service, agents can be particularly powerful, retrieving knowledge base articles, analyzing similar cases and generating email responses. After a case is closed, the AI agent can also summarize it, meaning that human agents and managers save considerable time on administration. By reducing repetitive and mundane tasks, AI can totally transform the experience for employees and for customers, operating within the same workflows people use every day.

The security risks of implementing AI

The costs begin rising as organizations offer up genAI access to more employees, which commonly happens once initial value is discovered. RHR, Greenberg said, has taken a cautious, security-first approach to genAI. An internal task force leads ongoing testing with a strong focus on privacy, ethics, and enterprise controls. Moving at the right pace helps target key workflows, avoid shadow IT, and build trust in the workforce. Leo CybSec unites a group of Cyber Security experts with 20+ years of collective expertise to help our clients realise and mitigate the cyber challenges and risks facing their business.

Enterprises should define clear visions and secure stakeholder buy-in to align AI initiatives with business goals. Centralising vendor management within unified platforms, like Kong’s API gateway, can streamline data pipelines and reduce complexity. The report suggests that enterprises are increasingly adopting multiple models – 37% use five or more – to optimise performance and avoid vendor lock-in. The findings, based on a survey of 550 IT leaders, developers, and engineers conducted between 26 February and 31 March 2025, highlight the rapid adoption of generative AI (GenAI) across industries. Another concern for Anthropic’s Askell is that AI tools can play with perceptions of reality in subtle ways, such as when offering factually incorrect or biased information as the truth. OpenAI said it is tweaking its training techniques to explicitly steer the model away from sycophancy while building more “guardrails” to protect against such responses.