Has AI Already Rewritten the Social Contract at Work?

For years, we have debated whether artificial intelligence will replace workers, but on reflection, have we been asking the wrong question? Should we consider whether AI has already rewritten the social contract at work? 

I think there is evidence to suggest that it has. The shift is not happening through humanoid robots replacing employees or dramatic announcements from boardrooms. It is happening quietly, one AI-powered hiring filter, one productivity dashboard, one performance score, and one algorithmic recommendation at a time.

Every organisation wants to become more data-driven, which increasingly means using AI to screen candidates, recommend promotions, analyse productivity, monitor communications, identify flight risks, score performance and influence who receives development opportunities. Before long, employees find themselves working for organisations that understand more about them than they understand about how decisions affecting their careers are made.

Why the Employment Contract Depends on Trust

The employment contract has never simply been an exchange of labour for salary. It has always rested on something less tangible but far more valuable: trust. Employees trust they will be assessed fairly; employers trust people to act responsibly; managers apply judgement because they understand context. In turn, workers accept difficult decisions because they believe another human has made them. AI changes those assumptions.

When algorithms quietly shape people’s decisions, trust must be intentionally built rather than assumed. The organisations that succeed over the next decade will not necessarily be those with the most sophisticated AI but those that earn what I call the Trust Dividend: the measurable enterprise value created when employees understand, influence and trust how AI is used in decisions that affect them.

When AI Objectivity Becomes a Risk

Many organisations are introducing AI into HR under the banner of objectivity. Yet AI is only as objective as the data, assumptions, and incentives behind it. A hiring model trained on decades of recruitment decisions may automate historical bias. A productivity score based on keyboard activity may reward visibility over value. A promotion algorithm may reinforce yesterday’s leadership model instead of identifying tomorrow’s.

Technology does not eliminate bias, however; it can industrialise it, and employees instinctively understand this. When people feel they are being constantly measured but cannot understand how those measurements influence decisions, or worse, the measurements negatively impact them, behaviour changes. Workers begin optimising for the algorithm rather than the organisation, with creativity making way for caution. 

Trust as a Competitive Advantage

The organisations that benefit most from AI will not automate the fastest. They will build trust the fastest. That matters because trust is already one of the most valuable assets an organisation possesses. Edelman’s Trust Barometer has consistently found that employees trust their employer more than governments, the media or business generally. As AI becomes increasingly influential in hiring, performance and career progression, organisations have far more to lose than operational efficiency if that trust is eroded. They risk damaging one of the strongest relationships they have with the people who create their value.

Employees should know where AI influences decisions affecting their careers, what data is being used and when human judgement can override an algorithm. Transparency cannot be reduced to an employee handbook or a compliance policy. It requires ongoing dialogue.

Participation Must Sit at the Centre of AI Deployment 

Treating AI deployment as purely a technology programme is a mistake. It is an organisational transformation programme. Employees should help shape the systems that influence their working lives. Representative employee groups should contribute to governance. New AI processes should be tested with those they affect, and clear appeal mechanisms should exist when people believe automated recommendations are wrong.

Some organisations already recognise this. Salesforce established an Office of Ethical and Humane Use years before generative AI became mainstream. Rather than viewing governance simply as regulatory compliance, it embedded transparency, accountability, and human oversight into AI development. The lesson is simple: employee confidence is not a by-product of responsible AI. It is a prerequisite for successful AI adoption.

Earning the Trust Dividend

The Trust Dividend is the return organisations receive when employees understand, influence and trust how AI is used. Trust accelerates adoption, encourages experimentation, and strengthens confidence that career decisions are fair, transparent, and accountable. Unlike productivity gains, which are often realised in months, the Trust Dividend compounds over years through higher engagement, stronger retention, faster innovation and greater organisational resilience. For too long, AI value has been measured through efficiency, automation, and cost reduction. Those metrics matter, but they are incomplete.

An AI system may reduce recruitment costs while damaging candidate confidence. Productivity dashboards may improve utilisation while employee engagement falls. Managers may become so dependent on algorithmic recommendations that they lose confidence in exercising their own judgement.

On paper, the AI initiative succeeded, but in reality, enterprise value declined. The next generation of AI measurement must extend beyond operational efficiency to include trust, transparency, fairness, employee confidence, adoption and organisational resilience. These are not soft metrics. They are leading indicators of an organisation’s ability to innovate, retain talent and create sustainable value.

This is where AI governance must evolve, as its purpose is not simply to satisfy regulators or reduce legal risk but to strengthen the relationship between organisations and the people who power them. Every AI system influencing people’s decisions should be explainable, contestable, and accountable, perhaps one of the greatest misconceptions surrounding AI.

If employees trust the systems around them, AI becomes an extraordinary force multiplier, helping people make better decisions and focus on work that creates genuine value. If they do not, every algorithm becomes another reason to disengage.

The Future of Work Depends on Earned Trust

The future of work will not be defined by how intelligent AI becomes. It will be defined by whether organisations earn and protect the trust of the people expected to work alongside it. Perhaps the greatest return on AI investment will not come from replacing human judgement, but from strengthening the confidence that allows people and intelligent systems to succeed together: the very definition of the Trust Dividend.