Most CEOs in the discussion are planning for AI to **augment work, not wipe out entire workforces**.

At the Semafor World Economy conference, Anthropic co-founder **Jack Clark** pushed back on predictions that AI will automatically drive unemployment as high as **20%** in the next five years. He framed large-scale job loss as partly a **policy and management choice**, not an inevitable outcome.

Key points for employees and leaders:

• AI is expected to **change how business is done**, from operations to national security to how teams collaborate.
• Leaders see AI as a way to **improve the quantity and quality of work**, not simply cut headcount.
• Any major labor market shift would **take time to play out**, giving organizations room to adapt through reskilling and role redesign.

In practice, executives on the panel are:

• Using AI to **get more done in the same 8–12 hour day**, not to reduce their own workload.
• Investing in **company-wide reskilling**, as Infosys is doing with its **300,000 employees**.

For most knowledge workers, the near-term reality is that AI will **reshape tasks and required skills** more than it will instantly eliminate roles. Organizations that plan for augmentation and reskilling are better positioned than those that treat AI purely as a cost-cutting tool.

Current AI usage is meaningful but still far from universal in daily workflows.

According to **Jon Clifton, CEO of Gallup**:

• About **50% of American employees** are using AI in some way.
• Only **13%** are using AI **on a daily basis**.

This gap explains why many organizations **aren’t yet seeing clear productivity gains** at scale. AI is present, but it’s often:

• Used occasionally rather than embedded in core processes.
• Adopted by individuals, not systematically rolled out across teams.

For companies, this suggests two priorities:

1. Move from experimentation to integration. Treat AI as part of standard workflows (for analysis, drafting, research, support), not just a side tool.
2. Support everyday usage. Provide training, guidelines, and clear use cases so more employees feel confident using AI **daily**, not just sporadically.

As daily usage grows beyond that **13%**, organizations are more likely to see measurable improvements in productivity and work quality.

Panelists highlighted two practical pillars of an effective AI strategy: **clear ownership** and **structured reskilling**.

1. Assign clear leadership for AI
Daniel Herscovici, president and CEO of Plume, emphasized the value of a dedicated AI leader:

• Plume has an **“AI czar”** who is responsible for defining and driving the company’s AI strategy.
• This role focuses on **how to implement the right infrastructure** and ensure AI is used in a coherent way across the business.

With this approach, Herscovici notes he’s **not working less**, but he is **getting more done** in his 8–12 hour day—AI is amplifying his output, not replacing his effort.

2. Invest in structured reskilling, not just tools
Salil Parekh, CEO of Infosys, described a company-wide reskilling program:

• Infosys is working to **re-skill all 300,000 employees on AI tools**.
• New graduates first spend **2–3 months learning software development without AI**, to build fundamentals.
• Only after that do they **introduce AI tools** to show how those fundamentals can be enhanced.

This staged approach helps employees:

• Understand the underlying work, not just the AI interface.
• Use AI to **enhance quality and speed**, rather than become dependent on it for basic thinking.

3. Focus skills on analysis and synthesis, not rote tasks
Jack Clark noted that workers entering the market should focus less on **rote programming** and more on:

• Analyzing and connecting information across **many disciplines**.
• Knowing **which questions to ask** and how to combine insights from different domains.

In short, an effective AI strategy:

• Has a **named owner** responsible for AI direction.
• Includes **company-wide training and reskilling**, not just tool deployment.
• Prepares employees to **reimagine their roles** around higher-value thinking, while AI handles more of the routine work.