Teradata Partners Conference 2017: the critical art of AI thinking

Thinking and talking about AI and data

Data quality and critical thinking are imperative for the advance of artificial intelligence (AI), according to expert views from an event hosted by US-based Teradata.

At day three (24 October) of the Teradata Partners Conference 2017 (22-26 October), in Anaheim, California, Banking Technology was there to gauge the zeitgeist in our data-drenched world.

As a quick review, on Monday (23 October), the company set out its plans in a super-slick presentation; unveiled Teradata Analytics Platform; revealed Danske Bank has launched its AI driven fraud detection platform; and showed off Agile Analytics Factory.

In a panel, “Should We All Learn to Code? The Era of Human-Machine Collaboration”, Stephen Brobst and Yasmeen Ahmad from Teradata; Nadeem Gulzar, Danske Bank; and Andrew Stephen, Said Business School – University of Oxford; discussed a lot more than just coding.

Stephen says data quality and critical thinking “should be a given but it’s not happening”. The issue is not to blindly accept what the AI provides. And as Ahmad points out: “Data with bias can potentially be fed in.”

Stephen thinks people won’t need to learn to code to interact AI, but there will be a “deeper appreciation of coding skills”. But as many people say at fintech events, he thinks there is a digital skills gap.

In terms of AI for data science, Ahmad sees the good as this automates the process and gives people more time for interpretation of models, so they can innovate the next generation of models. And from there the new ones can be applied, which in turn can free up data scientists from laborious tasks.

While the panel clearly had adoration of AI, this was not an all-consuming ardour you may see in an overwrought romance novel. Gulzar says AI will “help humans focus on what is relevant” but users should “question the model”.

The idea is to think about what the AI provides. An interesting idea, but I don’t see that happening. As an analogy, no one questions a calculator. You input, get the results and move on. AI will become like that. Who will have time to question everything? Few do it now when it comes to TV news, politicians, smartphones etc.

Brobst believes that a natural language interface is “the way to go but is not in arm’s reach yet”. However, he wants to see voice and text options. There will be quite a few occasions when typing in private will work better than revealing your intentions to a café full of people.

It was a wide-ranging discussion, not just related to fintech of course, with interesting views from all the panel.

We can expect to see, read and hear a lot more on the themes of data and AI. Which isn’t a bad thing because society needs to shape how AI develops to ensure most of humanity doesn’t lose out.

Avoiding madness

Elsewhere, in a presentation on “Mapping Encryption to GDPR requirements: practical use cases”, Mark Bower, global director, product management, Micro Focus, discussed the General Data Protection Regulation (GDPR).

Data time in Anaheim

Micro Focus is a UK-based software and information technology business, and with Bower’s 20 years’ experience of data he was well placed to offer some insight on this theme.

And so has Banking Technology recently – there was a feature on “Lessons from the GDPR compliance journey of a large financial services organisation” and opinions on getting ready for GDPR and Brexit.

There hasn’t been a shortage of coverage, but Bower’s views are welcome. Micro Focus spoke to firms, who say GDPR does not just create the risks of fines but also consumer litigation.

He also offered up some useful analysis – from the World Economic Forum – on citizens in the US and Germany, and their risk perceptions of GDPR.

Both nations shared similar anxieties, but German citizens wanted more transparency on which other companies could access their data.

Bower believes GDPR will be the same experience or “madness” as the Payment Card Industry Data Security Standard (PCI DSS) introduced to handle branded credit cards from the major card schemes.

This sequence of events comprised reaction; high costs; cost reduction strategy; change as PCI DSS evolves; and finally, a steady state. He urged the audience not to make the same mistakes as PCI.

In terms of data choices to cope with GDPR – these include deletion; anonymisation; pseudonymisation; and traditional encryption.

If you’re not clear on the meaning of pseudonymisation, this is a “procedure by which the most identifying fields within a data record are replaced by one or more artificial identifiers, or pseudonyms”. Bower states this retains logic and the meaning in data.

In answer to why not anonymise everything? Bower says this “requires 100% assurance and breaks the value and meaning in the data”. It also requires each data set to be analysed, and there are no standards to validate to.

With GDPR set to arrive in May 2018, we have not heard the last of this subject. But the views of data experts are worth paying attention to.

There will be more news from the Teradata Partners Conference 2017 (22-26 October) in Anaheim, California, as Banking Technology is here all week.