The AI complexity paradox: More productivity, more responsibilities

peepo/Getty Images

Does artificial intelligence (AI) make working life easier or complicated? Experts suggest the answer depends on the context. 

In a recent IDC-hosted interview, SIAC CEO Toni Townes-Whitley described AI as the ultimate weapon against system complexity, noting that her company is employing AI to reduce tech complexity in some of the most complex technology environments on the planet -- within the US Department of Defense. 

Also: Amazon's Andy Jassy says AI will take some jobs but make others more 'interesting'

Her team has been able to reduce mission planning and other operational functions at the department from "hours to minutes," she said. AI can have the same impact in commercial businesses, "reducing time and toil for business development, proposal development, searching, and creating new documents and content." On the developer side, AI has reduced the time spent on code generation.

These results are positive. However, other voices advised caution, as AI is just as capable of increasing as reducing complexity. The impact depends on where and how AI is applied, with the right kind of governance, of course. 

Also: You've heard about AI killing jobs, but here are 15 news ones AI could create

"The integration of AI into our technological landscape brings with it a host of new complexities," said Supriya Bachal, program manager for R&D at Siemens. 

"These complexities exist on multiple levels. Individual engineers and developers who are tasked with integrating AI into the tools we use must deal with new levels of complexity, as do the organizations that must manage these new AI systems."

Skill requirements may complicate the situation. While AI might potentially reduce the need for headcount in many areas, particularly in coding and IT management, applying the technology requires expertise in AI-friendly programming languages and frameworks, machine learning, deep learning, natural language processing (NLP), analytics, math, statistics, algorithm design, and deductive reasoning.  

"With AI-driven solutions across apps, APIs, and varied user endpoints, the IT landscape will become increasingly intricate," said Amitha Pulijala, vice president at Vonage. "This will require more specialized expertise to manage these new tools."

Also: Phishers built fake Okta and Microsoft 365 login sites with AI - here's how to protect yourself

At the same time, "AI has shifted the focus from foundational IT skills to use cases, implementation, and user experience," said Dennis Perpetua, senior VP of digital workplace services at Kyndryl. "This change opens up opportunities for new talent to use AI tools to accelerate their careers in IT."

It was suggested that open collaboration on AI initiatives is the key to overcoming skills challenges, an approach that brings together developers, data scientists, IT teams, and business stakeholders. 

Alleviating workplace challenges

When it comes to operational complexity, AI offers a mixed bag of benefits, but with many ways to overcome the issues. "AI can automate routine tasks, streamline processes, and in some cases, directly manage the intricate webs of applications and services that comprise our modern IT architectures," said Siemens' Bachal. 

"Tools like AI-driven observability platforms enable problem-solving in the digital space that would otherwise require a lot of human attention and cognitive load to simulate and synthesize."

Also: 5 ways you can plug the widening AI skills gap at your business

Another plus of emerging technology is that "when platforms run into operational problems, AI doesn't just offer infinite human problem-solving power; it can also do some problem-solving work of its own," Bachal added.

It was suggested that AI will hold the ability to help alleviate workplace challenges. "The technology can help optimize workflows, automate simple app development, and provide insights into system performance, freeing up developers and IT teams to focus on higher-value tasks," said Vonage's Pulijala.

In short, while AI is making technology access more complex in some circumstances, AI is also helping to manage complexity. "Overall, while AI has increased the complexity in certain aspects of IT, it has also brought significant efficiencies, creativity, and productivity, making the challenges worthwhile," said Kyndryl's Perpetua.

For example, he pointed out, "tools like GitHub Copilot are increasing efficiency in coding tasks, and AI-based APIs are becoming more autonomous, reducing the time spent on creating and maintaining them."

Also: AI won't take your job, but this definitely will

Another consideration is NLP, considered the gateway to the AI world, capable of breaking down traditional integration barriers between APIs and simplifying sprawling infrastructures, said Loren Absher, a director who leads ISG's AI advisory practice in the Americas. However, this NLP-enabled progress comes with challenges: "Machines interpreting human language must untangle ambiguous queries, maintain security, and ensure precision -- all while scaling dynamically."

He said that good governance is critical. "AI should be employed not just for automation -- such as monitoring, issue detection, and optimization -- but also as a mediator between traditional and NLP-enabled APIs," said Absher. "Tools like middleware platforms and orchestration engines can facilitate seamless communication across diverse systems."

Design AI systems "with transparency, adaptability, and robust security protocols," Absher advised. "A strong governance framework and ongoing investment in training and tools will ensure teams can harness AI's transformative power without losing control." 

Watch out for the agents

It was also suggested at the event that agentic AI could simplify rather than exacerbate complexity. 

"Agents can streamline ecosystems by connecting legacy applications, APIs, and disparate data sources," said Aaron Kesler, vice president at RozieAI, and formerly director of AI product management at SnapLogic. 

"They can identify inefficiencies, flag bottlenecks, and automate streamlined workflows optimized for existing systems, without requiring custom code or extensive dev time."

Also: What are AI agents? How to access a team of personalized assistants

For example, "fraud-detection agents can autonomously analyze transactions, flagging suspicious patterns while providing actionable insights to human analysts," said Kesler. 

"Similarly, research agents can scan the web to track mentions of specific products, aggregating data in real time to keep teams informed and proactive. All of this can now be built without heavy reliance on the data science team. Tasks can now be accomplished by one or two data engineers within the IT department."

Still, it's important to note that AI's impact on complexity will vary on a case-by-case basis. 

"For organizations with an already robust IT infrastructure and team, AI will probably just shift resources from one place to another," said Brandon Andersen, technology consultant and co-founder of Ceralytics. 

"Instead of a team working on maintaining legacy systems, they will now troubleshoot and periodically maintain the new AI systems -- especially the litany of connections."

Also: The most critical job skill you need to thrive in the AI revolution

For smaller IT teams, the story may be very different. "Instead of being an escalation point for current SaaS systems, these teams will now be in charge of the various API connections and be the first line of defense when the system goes down," Andersen said. 

"IT teams will absorb a lot of responsibility for these systems because the connections are no longer owned by a third party."