As we step into 2025, I find myself deeply contemplating the transformative potential of data and AI. At Intellinexus, we witness daily how these technologies are reshaping our world. Yet, amid the excitement, I'm increasingly mindful of our responsibility to harness these powerful tools thoughtfully and ethically, says Zimkhita Buwa, CEO of Intellinexus.
The promise and the reality
I'm both thrilled and in awe by the rapid advancement of AI technologies. It’s evolving at a pace that’s transforming how we connect and collaborate. The emergence of agentic AI particularly fascinates me, but it also keeps me acutely aware of its potential dangers. As my colleague, a senior manager at Intellinexus, Daniel Vlok recently reminded me: “AI agents will autonomously take actions, adapt in real-time and solve complex multi-step problems based on context and objectives.” While this opens up extraordinary possibilities, it also presents significant risks we cannot ignore.
In a recent interview with Forbes, Sam Altman, CEO of OpenAI, the company behind arguably the most famous generative AI interface – ChatGPT – said that technology companies were showcasing impressive advancements in multimodal AI capabilities in 2024. This is a leap into the future as it allows for models to process and "generate content across text, images and audio, thereby enhancing their applicability across diverse domains".
Multimodal AI adds another layer of sophistication to AI solutions and offers key benefits that include:
- Increasingly natural interactions closer to how humans process and share information.
- Multiple data sources provide a richer context for interactions.
- Cross-validation across different modalities.
- More intuitive and comprehensive interactions.
The key is to remember that there are risks. Multimodal AI can generate and manipulate realistic images, videos and audio, which makes it easier for bad actors to create convincing false content of executives, employees or products, potentially impact reputations and making it easier to commit fraud. These systems also require vast quantities of data, which can create additional attack surfaces for data breaches. Then there are concerns around bias as these systems can perpetuate or amplify existing biases across multiple modalities – an AI system processing both images and text may exhibit compound biases in how it represents or interprets different demographics, for example.
Which brings the conversation directly back to the importance of the AI-data connection.
The foundation of success: Data-centric development
I've learned – sometimes the hard way – that even the most sophisticated AI systems are only as good as their foundation. This brings me to a truth that I hold dear: data-centric development is not just a trend; it's a fundamental necessity.
Jessica Lu, one of our AI/ML Engineers, is passionate about not only data but also costume design – she’s an award-winning cosplayer. "Raw data is like fabric. While you could simply sew a tube and call it a garment, careful tailoring, cutting and stitching transform it into sophisticated, high-fashion pieces. Similarly, data requires processing to become truly useful. Its power lies in being transformed and integrated with other datasets. Jumping straight into AI without any preparatory data work is like attempting the finishing stitches before the fabric is even cut. Ultimately, the quality of the final product depends on the raw materials and the processes they undergo."
At Intellinexus, this has become the cornerstone of our AI development and innovation. We focus intensively on:
- Ensuring our training datasets are not just accurate, but truly representative of diverse populations.
- Implementing robust governance policies that guide every aspect of our data handling.
- Maintaining pristine data quality standards throughout our operations.
This commitment to data is critical to ensuring AI innovation thrives – as Monrico Basson, Senior Solution Architect at Intellinexus, points out: “Organisations want to tap into the data monetisation trend and create value from their assets by offering them up to others. This means that the data has to be pristine, accurate, relevant and secure, and that companies have access to an architecture that allows them to leverage this trend to its full potential.”
The Snowflake Data Exchange – one of the solutions prioritised by Intellinexus – helps companies to set up private and public data exchanges to share and monetise data seamlessly while the Snowflake Marketplace provides a ready-built ecosystem for companies to discover and purchase third-party data assets.
The human element: Our north star
But here's what keeps me up at night: in our rush to embrace AI's potential, we risk overlooking the human element. This isn't just about technology; it's about people. That's why I'm passionate about our engagements that go beyond technical implementations to address the ethical implications of AI usage.
The hype around AI is real, and I’ve witnessed firsthand how organisations often dive headfirst into AI adoption without fully understanding the implications. This rush can lead to unintended consequences, including ethical dilemmas and biases that can undermine trust. That’s why I am passionate about advocating for a more measured approach. We need to prioritise preparation and education, ensuring that every team member understands both the technical aspects of AI and its ethical ramifications.
We need to balance innovation with responsibility, excitement with careful consideration.
AI governance and ethics
The recent events in South Africa's legal sector and global banking industry raise serious ethical concerns about AI governance and societal impact. While JPMorgan's CEO Jamie Dimon envisions a future where AI could lead to "dramatic improvement in workers' quality of life" with "three-and-a-half days a week" work schedules, the projected loss of 200 000 banking jobs tells a different story. The South African legal case, where AI was allegedly misused to generate false citations, highlights the urgent need for responsible AI implementation and proper training.
Rather than focusing solely on the estimated $180 billion in additional profits, we must consider our moral obligation to workers. While "any jobs involving routine, repetitive tasks are at risk", as analyst Tomasz Noetzel notes, the focus should be on workforce transformation rather than wholesale job elimination. The banking industry has an opportunity – and arguably a responsibility – to reinvest some of these AI-driven gains into comprehensive reskilling programmes for affected employees, particularly in back office and operations roles most at risk. This would help ensure AI advancement benefits society as a whole, not just corporate bottom lines.
Embracing the future
As I reflect on these developments, I'm reminded of why I entered this field: the potential to make a positive impact on society. At Intellinexus, we believe in lifting others as we rise. This means empowering diverse voices in technology and ensuring our innovations serve everyone equitably.
The future of AI is not just about technological advancement; it's about responsible progress that benefits humanity as a whole. As we navigate this exciting future, let's stay grounded in our values and committed to making a difference.
It is time to undertake AI from the perspective of responsible innovation – together, we can ensure that our advancements in data and AI create a future we're proud to pass on to the next generation.
Share