Computicket has announced the launch of its new self-service platform Box Office that lets organizers of small events sell tickets. The launch of the…
With the market value of Artificial Intelligence in fintech estimated to hit $26 billion, the computational arms race is on. And that’s just in fintech where the latest innovations are in rapid customer evaluation to improve access to finance.
South African data science specialists NGA has been a market leader in risk assessment through its RiskSecure anti-money laundering solution for the finance industry, making company CEO Mark Germishuys a great source of information for AI-related decision making.
“When NGA first started playing with data, we made a decision to ensure that our models used ethical AI,” he explains.
“We started small, so we know that most SMMEs can’t afford to do big data research. That’s why we built one of the biggest AI services on the continent.”
Companies as big as Goldman Sachs, Google and Amazon have faced financial and reputational damage when obvious bias was discovered in the AI tools used for credit allocation and hiring.
Ethical AI seeks to eliminate bias and discrimination by constantly updating and improving the data sets used in machine learning.
“Data needs controls because there’s more to it than simply scraping data,” Germishuys continues. “One of the biggest problems we see globally is false positives. Good AI is trained to look at context because it analyses sentiment and negative keywords. This helps deliver a clean data set that is free of the noise.”
South Africa’s Protection of Personal Information Act (POPIA) and similar data privacy regulations around the world have significantly impacted the way companies collect, process and store customer data.
“Modern AI can predict market trends because we have the entire history of the internet to trawl for patterns. And these patterns do repeat themselves,” he says.
“We also have the capacity to support our products because we’ve built them from scratch using revenue we generated from other parts of the business and not impatient venture capital.”
Data anonymity has become important as data privacy is becoming more regulated from a government level
That product support component is also linked to liability and the appetite for carrying the risks of increased scrutiny in the AI market. It’s also a major reason why an SME may opt to buy an of-the-shelf solution rather than build one from scratch.
Germishuys believes in the value of a slow build and not overpromising capabilities. “Don’t lie to clients about readiness, just get the MVP (minimum viable product) out,” he says, expressing his frustration with industry trends that create negative perceptions of AI technology.
Three solutions AI can create for any enterprise:
- Speed up research and development: AI solutions like SocialListener are intelligent enough to understand sentiment, which is a more accurate way of viewing trends. Because they analyse the language in the articles and social media posts related to your search (in 110 languages), your results reflect the true feeling around your next innovation.
- Marketing secrets revealed: Machine Learning can build a detailed map of product launch strategies from the biggest brands in the world and you can plug those findings directly into your strategy.
- Scale to your needs: Machine learning algorithms are only as good as the data it can train on. The other end of that equation means that they can adapt to the specific data related to your business, be it sales trends or the effect of next week’s weather on your crop.
Featured image by Kevin Ku/Unsplash