Innovate UK Transformative Technologies Grant
AI powered training and assessment of individuals

1 Jan 2024 – Artificial Intelligence holds enormous promise for accelerating innovation across industries. Yet, even cutting-edge solutions like ChatGPT have demonstrated the potential pitfalls of bias creeping into AI outputs—ranging from subtle gender assumptions to outright prejudices. Recognising these risks, Innovate UK has awarded a grant led by Dr. Mahdi Jelodari to develop an AI-powered simulation platform that systematically trains and assesses IT engineers in a bias-free environment. Below, we explore why addressing bias is critical, how simulation-based training can empower IT professionals, and the broader implications of this pioneering project.

The Reality of AI Bias

In February, Sam Altman, CEO of OpenAI, pointed out that ChatGPT has “shortcomings around bias.” Whether it’s describing a “good employee” as a white male or defaulting to male pronouns when describing a “smart software engineer,” these tendencies highlight a deeper issue with AI training. When models are built on vast internet datasets—which reflect societal imbalances and stereotypes—they run the risk of perpetuating those biases in their output.

This problem becomes especially urgent in professional contexts. Biased AI can influence hiring decisions, workplace culture, and even performance assessments. Moreover, employees who feel stereotyped may disengage, leading to reduced morale and productivity. In short, unchecked AI bias undermines both fairness and effectiveness in the workplace.

Why Bias-Free Training Matters

A bias-free environment is not just about social good—it has tangible benefits for businesses and individuals alike:

• Increased Innovation: Diverse viewpoints foster creativity and problem-solving. When AI treats individuals fairly, it empowers every engineer to contribute unique ideas.

• Higher Employee Engagement: Employees who feel respected and valued are more engaged and productive. Reducing bias is key to creating inclusive work cultures.

• Enhanced Collaboration: Training that actively includes diverse scenarios helps teams learn to communicate effectively under pressure, building cohesion and trust.

• Reduced Turnover and Absenteeism: Fair, unbiased work environments increase employee satisfaction, boosting retention and overall organizational health.

Unleashing the Power of Simulation

Investing in simulation training can significantly improve how quickly IT teams identify and resolve outages. Real-world incidents often incur high costs—from lost revenue to reputational damage. By practicing in a virtual yet realistic setting, teams can:

• Develop Rapid Response Skills: Hands-on practice helps engineers detect and fix problems more quickly when real crises hit.

• Minimise Economic Impact: Reducing downtime and preventing mistakes can save businesses millions.

• Protect Mental Health: Simulation-based learning lets teams make (and learn from) mistakes safely—without the pressure and stress of a live environment.

Moving From Supervised to Unsupervised Training

While expert supervision remains key in early-stage simulations, the ultimate goal is to automate the platform so that it becomes widely accessible. This transition requires “baking in” a wealth of domain-specific knowledge into AI models, ensuring they remain both powerful and free from bias.

Over the course of seven months, the Innovate UK grant enabled Dr Mahdi Jelodari to hire additional experts and gather comprehensive simulation data. This data helped further refine the AI algorithms, creating robust, accurate, and fair models. The result is a fully autonomous training platform that can be seamlessly adopted by organisations across various industries.

A New Standard in Tech Training

Thanks to the leadership of Dr. Mahdi Jelodari and support from Innovate UK, Uptime Labs is on the path to redefining how IT professionals are trained—by combining the rigor of simulation with AI-driven insights. By tackling the bias problem head-on, the platform ensures equal opportunities for learning, growth, and recognition, helping every engineer realize their full potential.

Key Takeaways:

• AI Bias is Real: Even advanced models can inadvertently perpetuate stereotypes.

• Simulation is Essential: Safe, realistic training boosts confidence and crisis-readiness.

• Expert Supervision Prevents Bias: Reviewing AI outputs ensures fairness and psychological safety.

• Automation = Scalability: Once bias-free AI models are established, the platform can train vast numbers of engineers quickly and cost-effectively.

Looking Ahead

Unbiased AI isn’t just a technical challenge; it’s a societal imperative. As we refine our simulation platform, we strive to create an environment where all IT engineers can thrive—regardless of gender, ethnicity, or background. By demonstrating how AI can be harnessed responsibly, we hope to inspire broader conversations and innovations that make technology work for everyone.

Authored by Dr. Mahdi Jelodari

Link to the grant