Rikus Combrinck teaches machines how to collect and process data like humans, but more efficiently. In 1982, growing up in Prieska, South Africa, he encountered a VIC-20 computer during a school outing to Kimberly. It had 5 kilobytes of memory, much less than required by a single desktop icon on current computers. It used a television for display and a tape drive for storage. As a young boy he loved disassembling things, exercising his analytical mind. After studying electronic engineering, Rikus’s postgraduate studies included AI and machine learning, which fascinated him.
After many years in software development he reconnected to his early interests and now practices as a data scientist, helping organisations analyse their data to detect patterns that can be used to solve business problems. Current projects include the accurate valuation of used vehicles for getWorth (getworth.co.za) and automating the marking of adult maturity assessments for Aephoria (aephoria.co.za). Data-centric organisations like these invest in AI because decision making based on accurate data clearly gives a competitive edge. Heads up programmers, the industry is growing rapidly as organisations realise that these valuable tools are becoming more accessible.
Machines are getting better at making decisions like humans, but still have a long way to go. For example, with the Aephoria’s maturity assessments, machines struggle to process subtle incongruences or ironic nuances in questionnaire responses. But the ability to process human language is rapidly developing. To help solve linguistic problems Rikus uses AI tools made freely available to the public by companies like OpenAI (founded by Elon Musk), Google and Facebook. The latest version of OpenAI’s Generative Pre-trained Transformer (GPT-3), an autoregressive language model that uses deep learning to produce human-like text, has not been released for public use as it is deemed unsafe. The quality of the text it produces is frequently indistinguishable from human writing, posing risks that outweigh the benefits. In the wrong hands it could be used to impersonate users communicating online for criminal gain.
Rikus acknowledges the fear of computers replacing us in the workplace, along with its social and economic complexities. However, as a result of AI becoming more pervasive, products and services will become affordable to people who do not currently have access to them. This includes critical services like healthcare and education. So the shrinking job market may be offset by a decreased need to earn as much as we do now. This is a generalisation but the idea is that if all goes well, AI could actually serve us whilst tasks that require a personal touch will remain in our hands.
Rikus speaks of an online backlash against freebie-based advertising, adding that Google is in the process of eliminating third-party cookies in favour of less invasive ways to gather browsing statistics across large groups. People simply won’t stand for surreptitious and unsolicited marketing practices for much longer. Although lack of privacy is mainly a threat to those who have something to hide, there are also security issues at play. As time passes and technology matures our concerns will hopefully be better addressed. In Rikus’s opinion we are already cyborgs as information technology is seamlessly integral to our daily functioning.
For more thought pieces by Dimitri, please visit www.dimitriotis.com.
If you are looking for a trusted solution to help your organisation orchestrate and analyse your data, mitigate risks, increase productivity, improve efficiencies and increase your bottom line results, reach out to us by emailing firstname.lastname@example.org and one of experts will get in touch to schedule a demo and discuss a tailor-made solution.