New IMD research 2025
Mind the Gap
Addressing the risk of bias in Generative AI
Accelerating integration of Generative AI (GenAI) will drive massive productivity gains, potentially adding trillions to the global economy. But even as organizations look to leverage its promise, the imperative to address bias in training data, algorithms and outputs is becoming more urgent.
IMD has launched a white paper together with Microsoft Switzerland and the Ringier AG EqualVoice Initiative that sheds light on:
- How bias manifests in the GenAI lifecycle.
- The risk of bias in GenAI for organizations and society at large.
- Effective governance frameworks and mechanisms to help organizations identify and tackle bias in GenAI wherever it surfaces.
- This white paper also shares the findings of a proprietary global survey of organizations and leaders.
Bias in GenAI data, algorithms and outputs
Bias enters GenAI via training data that mirrors real-world societal inequities and discrimination.
A lack of diversity in design, building and testing processes can then make it hard for organizations to identify bias in their development and deployment processes.
Once biased outputs are deployed and make their way back into the real world, they have the potential to perpetuate stereotypes, widening existing gaps in fairness, representation and access – and at scale.
How biased is AI?
AI prompt results
Research by EqualVoice (November 2024) reveals bias in prompt results.
Bias in GenAI: a risk to people, organizations and society
Bias harms those who are discriminated against but also constitutes a threat to organizations.
If businesses are using biased outputs to help map markets, design products or engage with user bases they run the risk of skewed decision making, missed opportunities, and damage to their brand or reputation and customer loyalty.
As integration accelerates, so too will regulation, meaning that organizations that fail to comply are more likely to face financial and operational consequences.
How did survey respondents rank the risks of bias in GenAI for organizations?
Organizations are underequipped to tackle bias in GenAI
The majority of global survey respondents said they are optimistic that GenAI will transform productivity and help drive innovation.
But 72% of respondents are concerned about bias in AI and many don’t feel equipped to address the problem.
A pan-organizational approach to minimizing bias in GenAI
Tackling bias in GenAI starts with determining responsible AI principles that tie to organizational values.
These principles must then be enacted through effective governance frameworks, cross-functional collaboration and a culture of shared accountability.
This means de-siloing the organization to leverage diverse knowledge, skills, perspectives and capabilities, and looking to the broader ecosystem.
Diversity, equity and inclusion functions, inclusive leadership, and ongoing training and education have critical roles to play.
The white paper shares insights and recommendations that focus on three spokes
Processes
Technology
People
Learn with IMD
Prepare your organization’s digital future
How do you upskill to stay ahead in the digital age?
The nature of work is changing, and fast. IMD’s rich portfolio of digital transformation and AI programs equips you with the right skills, useful frameworks, exclusive research and case studies to understand data analytics, harness the power of AI and rethink your strategy through the lens of digital. Stay relevant and become a digital transformation expert with IMD.
Explore more
TONOMUS Global Center for Digital and AI Transformation
Helping leaders navigate digital and AI transformation journeys
Diversity, Equity & Inclusion Initiative
Helping leaders and organizations to invest in inclusive leadership, equity, and fairness
I by IMD
Bias in Generative AI: A risk that must be addressed now
As GenAI accelerates towards near universal integration, bias in data, algorithms and outcomes can be a significant risk to businesses and society. An IMD white paper with Microsoft and the Ringer AG EqualVoice Initiative, launched at Davos, shares proprietary survey insights and recommendations.