Metal,
Not Magic.

Susi O'Neill wrote a book about inclusive AI for a UK publisher.

She had years of research and a contract. Then the political climate in the United States shifted and the publisher came back with a request — change the title, soften the framing. Susi said no. Her book was canceled.

What you might do at that point is shelve the project.

What Susi did was take the entire book — the research and the resource pack — and release it for free. Anyone who scans her QR code receives the deck and a bundle of materials they can use in their own work.

The Social Saturday session on May 9, 2026 was the first time she had ever delivered this talk publicly. She came out of the chute with a keynote that holds together as a single argument and ends with a menu of commitments she asked the audience to choose from.

Her book got canceled. The argument did not.

She showed up at Social Saturday with that line as the anchor of the whole project and led us through what it means and what an inclusive AI rollout looks like in practice. The She Leads AI community came away with a free toolkit and a sharper way to talk about AI at work.

There is a story circulating in the boardroom that AI adoption is following a normal technology curve and the laggards will eventually catch up. Susi sees something different.

Generative AI is foundational. It behaves differently than the data-led and structural technologies that came before it. It reshapes how value is created and how careers develop. Whether you welcome it or not, the shift is already underway. Lagging at the early-adoption stage compounds.

So the question is — who is being left behind, and why?

The data Susi compiled across her annual International Women's Day research tells a consistent story. Women at work are about 25% less likely than men to use AI, per Harvard Business Review. Workers over 50 are less likely to engage. The Edelman Trust Barometer shows trust in AI and AI companies dropping fast, with the steepest declines among women and older workers in Europe and North America.

Only about half of all workers receive any formal AI training. The half that does shows exponential capability growth. The half that doesn't is being told to manage it on their own.

The skills gap is the surface. Underneath sits an eroded trust crisis and a rationed training pipeline, with historic bias the algorithms now reproduce at scale.

Two points from her data carry the argument —

Women in equivalent senior-leadership positions are still about 16% less likely to use AI than men in the same roles. Genevieve Smith and her husband had matching age and matching income — but her credit limit came in lower because gender was the only factor.

He could have been talking about AI adoption. The further away you are from being a Silicon Valley bro, the less likely you are to experience the benefits.

Susi's framework identifies the places where AI adoption goes sideways for women and other underrepresented groups. Each is a place where a leader or a team can intervene.

Susi's argument turns at the line AI is metal, not magic. Inclusive AI is small wins measured over time, with training and psychological safety to make them stick. Her framework defines four principles for leaders rolling out AI programs that lift the whole workforce.