EP13: Tech for Good: Empathy, Ethics & Youth in AI with OneHive

Personalising Learning at Scale

OneHive's platform is built around three core types of learning: consuming content, learning from a mentor or expert, and learning through trial and error. Rushab explains that the platform blends all three into a tailored experience for each individual — recognising that generic training, delivered the same way to everyone regardless of role or interest, struggles to drive real engagement. A memorable example: rather than teaching contract law generically, connecting the topic to something a learner is genuinely interested in — like Lewis Hamilton's Ferrari contract for a Formula 1 fan — creates far deeper engagement than a one-size-fits-all approach.

The AI Behind the Platform

Underpinning OneHive's personalisation is an AI matching algorithm that uses clustering methodology to group people based on shared experiential, educational, and personal attributes — automating a process that would otherwise be highly manual. The platform also addresses a common misconception about learning content: the challenge isn't a shortage of material, but information overload — helping the right person find the most relevant content for their specific needs, from vast existing content banks.

Rather than relying on traditional dashboards and manual data exports, OneHive enables organisations to query their own data conversationally — for example, asking which cohort members are most likely to disengage — turning what would otherwise require significant technical setup into something genuinely accessible for organisations without in-house engineering talent, a common constraint in the charity and public sectors.

Case Study: Scaling Impact at Tutors United

Rushab shares a detailed example involving Tutors United, a charity providing tutoring to thousands of children in foster care each month. Before working with OneHive, one staff member was making up to 200 calls a week to parents — chasing session links, checking on homework, following up on attendance — creating serious key-person dependency risk. OneHive's approach began with a full assessment of the charity's existing systems and bottlenecks, producing a clear, upfront "digital impact assessment" outlining expected outcomes — such as an 80% reduction in admin, or a 20% increase in engagement — making it significantly easier for funders to justify investment.

Once deployed, automation (including automated invoicing) freed staff capacity, and the system now identifies the roughly 50 families most at risk of disengagement each week, allowing staff to focus their attention where it's needed most — rather than spreading thin across 200 calls regardless of individual need.

Addressing Risk: What Happens If You Don't Adopt AI?

When it comes to reassuring risk-conscious organisations like schools, charities, and local authorities, Rushab reframes the conversation: the biggest risk facing many of these organisations often isn't AI adoption itself, but underlying challenges like funding sustainability and staff turnover. On the practical side, he highlights how AI can actively strengthen safeguarding — for example, automatically monitoring messaging between adults and vulnerable young people to block and flag attempts to share personal contact details — turning AI into a tool for risk mitigation, not just a source of new risk, alongside standard data security and compliance measures like GDPR and ISO standards.

Bridging the Digital Divide for Young People

OneHive's roots trace back to Rushab's earlier work running a programme connecting volunteers with children with disabilities, pairing people based on shared interests to build confidence and engagement — a model that now underpins the platform's broader approach to youth engagement.

A standout example is a tech boot camp with the Young Brent Foundation, delivered alongside major organisations including Sky, KPMG, and PwC. Rather than sticking to a generic AI curriculum, OneHive adapted the programme based on participants' actual interests — bringing in a music AI company after discovering many participants were interested in music. That session, Rushab notes, proved far more effective at building genuine understanding of AI than a standard, generalised explanation — delivered to young people in a community centre, not a private school.

Rushab is clear that being "digitally native" doesn't automatically translate into technical career readiness — using social media doesn't equate to data science skills, just as playing video games doesn't equate to blockchain development skills. The goal, he argues, is helping young people from all backgrounds build genuine confidence and capability with AI, while also being taught to navigate risks like deepfakes and misinformation — treating AI literacy as something to embrace responsibly, not fear.

Tech for Good: Building Responsibility Into AI by Design

For Rushab, "tech for good" means embedding consideration of potential harms — such as how an algorithm might hallucinate, and who could be affected — into the design process from the outset, not as an afterthought. He pushes back on narratives painting technology as inherently harmful, arguing instead that responsible AI development, and genuine investment from organisations in social impact, are what will define the industry's legacy — comparing it to lessons learned from social media's rapid, under-considered adoption two decades ago.

The Business Case for Responsible AI

Both Rushab and Tim agree that responsible AI isn't just an ethical consideration — it's increasingly a commercial one. Consumers, particularly younger generations entering the workforce, are placing growing value on organisations that demonstrably act responsibly, making genuine commitment to tech for good a meaningful competitive advantage rather than a compliance box-tick.