Interview: Rackspace Technology’s backing for Straits Interactive in AI assistant development
In the second part of our interview with Rackspace Technology and Straits Interactive, we get some insights from Hemanta Banerjee VP, Public Cloud Data Services, Rackspace Technology.
Hemanta was asked about the company’s role and support to Straits Interactive in the development of the AI DPO assistant.
He says the leading end-to-end hybrid multi-cloud technology services from Rackspace helped Straits Interactive design, build and operate the AI DPO Assistant as part of the Capabara Capability-as-a-Service platform.
The company’s secure multi-cloud solutions and Microsoft Azure’s OpenAI Service also ensured that Straits Interactive could help their customers accelerate generative AI adoption while balancing risks, safeguarding data, and maintaining compliance.
Rackspace’s FAIR
“For example, through our IP in the Foundry for AI called FAIR offering, we gathered specific requirements from Straits Interactive to propose a multi-phase approach.
“This facilitated a prompt demo for stakeholders, which met short-term needs, while ensuring a scalable, productised solution for long-term customer needs. By implementing a responsible OpenAI framework alongside IaC and CI/CD frameworks, Rackspace ensured faster iterations which optimised efficiency and effectiveness,” he says.
Hemanta also spoke on the challenges in the adoption of Responsible AI.
He says one of the biggest concerns relates to clarifying which information is shared with underlying platforms for services that integrate AI functionality. This is due to AI systems inherently having multiple layers of information encoded as part of the models deployed.
“Arguably the most talked about issue is with regard to protecting user privacy while training AI systems. While no small feat, this can be achieved through models being built to balance data access while ensuring privacy.”
Meanwhile, how AI algorithms perpetuate biases that lead to less impactful outcomes or even further entrench inequalities, is also hotly discussed. Expect this to drive action on ethical usage, creating space for policy and solutions in areas such as Accountability, Reliability, Explanability, Security and Privacy, he says.
AI in the next 5 years
“It would be bold to say what exactly AI will look in the next five years because it is evolving incredibly quickly. That being said, it’s not far fetched to imagine it will continue to be democratised – especially when you look at how generative AI has been used. Cloud computing, widely available pre-trained models, and the open source nature of most developments will usher in accelerated use setting up 2024 as a firm foundation for wider adoption in the next half decade or so. The use of natural language models will also broaden the adoption of AI.
“The development of AI will also likely have a lasting impact on sustainability. For instance, widespread usage could drive up emissions, as expansive use of compute/storage requires massive amounts of energy to power and cool. It’s worth noting, however, that this is being met with organic movements across the tech industry to drive sustainability in every aspect of technology. Furthermore, AI technologies can also be leveraged to pinpoint where wastage or unnecessarily high consumption occurs, which can then be corrected,” he concludes.
Table of Contents
Random pictures
