Red Hat’s Beatriz Ortega Eguizábal, stated, “We are in amazing momentum” – although at this stage 78% of algorithms fail to make it into production, there are pragmatic ways to succeed
Beatriz Ortega Eguizábal, EMEA Telco Business Development Manager at Red Hat presented To infinity and beyond – Open source AI unleashed at Mobile Europe’s virtual Telco to techco conference in April. As she explained, the purpose of her talk was two-fold. First to ensure that everyone is “on the same page” with AI in terms of identifying the opportunities, gaps and challenges that telcos face.
Second, to talk about why Red Hat, the leading open source vendor in telecoms, is talking about AI and its value proposition.
WATCH THE VIDEO NOW HERE
She points out that although GenAI is centre stage and generating data. Yet it lacks maturity, is subject to issues like hallucinations (which typically arise from insufficient training data, biases in data or the model making wrong assumptions) and at the height of the Gartner hype curve.
She noted, “We need to be sure that the content that is generated is real content is not fake news. That means that really the algorithms of the model are only one small piece in this ecosystem. So the data verification, the data collection, the configuration [and] the monitoring – the process – is one huge ecosystem that we need to control. Otherwise we will work in a lot of algorithms that are really good algorithms but when they go into production, they are failing.
“We can see that…around 78% of the models of use cases generated with AI never go to production because they fail in the in the production process.”
WATCH THE VIDEO NOW HERE to explore other issues and a highly pragmatic guide to where to start with deploying GenAI and the importance of Agentic AI in these early stages, and automation overall.
