When AI Agents Become the Workforce: The Future of Work

Software has traditionally been a passive tool. However, this is changing. We’re seeing a new category of software known as AI agents. These systems can set goals, plan steps to achieve these goals, utilise tools and learn from feedback.

As these systems become more robust, they can take on roles within organisations. This will fundamentally alter how we structure our employment, organisations and economy more than the introduction of the internet or spreadsheets.

They’re Not Just Faster Workers

This fundamentally changes how we use software. Whilst we still have a human guiding generative AI tools, we now have systems that can autonomously watch pipelines, negotiate standard contracts, deal with exceptions in inventory and manage processes across different platforms. We’re already seeing these systems take on processes requiring junior employees or specialty staff.

This will dramatically improve efficiency across processes. This software can now operate continuously across time zones, increasing productivity and reducing fixed costs that would have been spent employing staff. So we’re not just automating tasks, we’re introducing a whole new workforce that doesn’t require sleep, benefits or can be cloned with ease.

Human Labour Will Be Redefined

History reveals that certain occupations will be eliminated due to the adoption of modern technologies. However, the process is usually complicated. AI systems keep taking over massive routine intellectual activity volumes, which include data reconciliation, basic research preparation, report writing and tier one support.

People will be used for more strategic tasks. Fewer employees will be required for the processes mentioned earlier but more for the development, coordination and integration of the systems. The skills will be quite highly valued among those who can have an unambiguous goal in mind, do the critical analysis of the outputs and interact with the systems during edge cases.

More Services Will Become Cheaper

With the labour becoming cheap and plentiful, many services will become cheaper to provide. This will drive demand for many services such as healthcare management, personal learning, legal document review and software support. However, those whose skills can be replicated by AI agents will see their negotiating power decrease.

Recommended for You:

• Physical AI: Why It Is the Next Great Frontier in Technology
• Playground Global Returns to Taiwan: AI Hardware and Innovation
• How AI Is Redefining Human-Machine Interaction

Whilst there may not be widespread unemployment, we’ll see a greater divide between those who own/manipulate AI agents. If we don’t choose to implement policies that ensure the workforce are taxed, have portable benefits or invest in skills, those who own capital will benefit.

Those Who Adapt Workflows Will Benefit Most

Those who will benefit most from AI agents will be those who fully embed these systems into new workflows that utilise both human and AI agents. Those who can provide clear accountability, auditing and override will ensure these systems improve reliability, however those who don’t will introduce new risks.

Managers will need to learn how to manage these non-human workers. Many will struggle as many organisations still struggle to implement well understood digital tools. Those who see these agents as a miracle and don’t manage them will incur costs.

Liability & Ownership Issues

As the technology develops faster than our ability to manage it, we’ll quickly run into issues. If an AI agent makes a mistake with serious consequences, who is liable? If agents create work, who owns the IP? If agents are continuously learning from data within the organisation, who owns the data? Many laws regarding labour, licensing and health and safety have been created for human beings.

We’ll need to consider if the AI agent is simply a tool or has some level of autonomy. If societies fail to move quickly enough, we may fail to utilise these tools effectively. However, if we move too quickly without considering the risks, we risk a few providers having too much control.

There Is Still Time To Adapt

It is not too late for us to adjust. Individuals could concentrate their efforts on developing abilities that complement those of the AI agents. These include detailed knowledge in particular domains that would help guide the AI agents, information integration abilities, judgement and design of complex systems.

Education systems must abandon conventional methods of teaching and switch to methods of critical thinking, management and ethics. The governments can speed up the process through providing fast training, changing our social security systems in order to enable mobility among positions and securing energy, data and computing capabilities.

AI agents are becoming a real part of our workforce. This can be seen in many niche areas and will become more prevalent as they become more reliable and easier to integrate. How this affects us will depend more on how we choose to use them. Those who see this as an inevitability but can shape how they are used will be in a better position.