A Framework for Measuring Ethical AI

Wanted: A "2-Degree Target" for the AI Era - how can we track responsible AI?

This is part of a new series called #FutureWork that is trying to understand the rapidly changing workforce.

On yesterday’s flight to DC, I was reviewing economic projections and historical data (yes, I’m a geek), I had a realisation that haunts me: We're racing toward a tipping point that could fundamentally reshape society, yet we don’t know how fast we’re going. If we’re going to have a harmonious relationship with AI we need our generation's equivalent of the Paris Agreement's 2-degree target.

Let me explain what I mean.

How I understand the impact AI will have on the economy

I’ve been curious about how AI will affect the workforce and have been reading (2024 Nobel Prize Winning Economist) Daron Acemoglu's "Power and Progress," which tracks how each successive technological leap has affected the economy, in particular worker vs capital owner share of GDP. I have been trying to answer a critical question.“At what point does automation advance to such an extent that it undermines wage levels, jeopardizing the very consumption that drives economic growth?”

The numbers tell a pretty compelling story. Each technological revolution—PCs, the internet, mobile— decreases workers’ share of GDP. This leads to less of the economic pie and has coincided with increased wealth inequality.

According to OpenAI’s latest economic blueprint¹, we started at 65% in the pre-PC era. Today, we’re at just 57%. By 2030, projections from the World Economic Forum² show us hitting 52%—a tipping point where diminished consumer spending power could ignite an economic spiral that impacts everyone.

There is a threshold for workers’ share of GDP below which our current economic framework could unravel entirely. We are on course to hit it in 2030.

Why This Time Feels Different

Having led teams through multiple technology transitions, I've seen both the opportunity as well as the inevitable human cost. AI's capacity to impact 40% of working hours (Accenture, 2023) feels different to before. Digital technologies could give us a 2x improvement, but a 10x improvement using AI is achievable already across knowledge roles both vertically (your role is fully automated) and horizontally (we take away email and phone management, saving you lots of time). We’ve finally got the tools to unleash incredible progress, but if we don’t deploy them carefully I’m afraid we might inadvertently crash our economy.

The spiral I see forming keeps me awake:

- AI drives efficiency but displaces jobs faster than we create them

- Consumer spending power declines

- Demand drops

- Businesses implement more automation to maintain margins

- The cycle deepens

This isn’t new. Every technological revolution has presented similar challenges, and humanity has found ways to create new opportunities alongside new efficiencies, the question is… how quickly?. Research from Harvard Business School⁴ shows that when properly implemented, AI can increase both productivity and quality by over 40% while creating new kinds of work. So it’s possible. How do we measure it.

The Creation-Displacement Ratio (CDR)

I believe we can measure this, manage this, and master this at a local level (i.e. within individual teams and companies) - but only if we act now.

I propose every organization adopting AI commit to maintaining a "Creation-Displacement Ratio" (CDR) of at least 1:1.

Think of CDR as your company's ethical AI compass:

1. Track - how much of your headcount budget are you saving as a result of implementing AI tools.

2. Match: Equal investment in new job creation, skill development, or wage increases so that your employees can share in the rewards.

3. Measure: Your CDR measure is your company’s contribution to ensuring we don’t crash the economy.

Making It Real: The Corporate Pledge

Imagine if every company made this commitment:

"We pledge to maintain a minimum 1:1 Creation-Displacement Ratio through 2030, ensuring our AI implementation contributes to economic stability by preserving worker share of GDP above 57%."

This isn't about slowing innovation - it's about sustainable progress. DeepSeek shows us that irrespective of export controls, the U.S. government cannot slow down AI advancement. But there is one thing we all agree on:

Each of us hopes AI will unlock human potential. No one wants AI at the cost of chaos. This is possible, it is within our control.

Do you think you’d be open to signing a pledge like this? Or do you have a better idea? Let’s figure it out together.

How are you thinking about this?

References:

¹ OpenAI Economic Blueprint (2025)

² World Economic Forum "The Future of Jobs Report" (2023)

³ Accenture "GenAI LLM" Report (2023)

⁴ Dell'Acqua et al., "Navigating the Jagged Technological Frontier" (2023)

#AIFuture #EconomicInclusion #ResponsibleInnovation #Leadership