Avatar

This blog comes from Dr. Catriona Wallace is the Founder & CEO of Ethical AI Advisory and Executive Director of Flamingo Ai, provider of Machine Learning-based technologies. Catriona is one of the world’s most cited experts and speakers on Artificial Intelligence, customer experience, ethics, human rights in technology, and women in leadership. She recently spoke at a Women Rock-IT event, to hear more from Dr. Wallace, click here to view the video on demand. 

The continued development of Artificial Intelligence (AI) and Machine Learning (ML) will soon deliver hyper-personalisation, which is a supercharged type of analytics platform that serves up highly specific offers and content to customers. For example, if you are researching online, then you will be served ads specific to that holiday in applications like Facebook, email, and other browsers. Now, these personalisation engines get it about 40 percent right. Essentially, they capture the shadow of the person on the internet, but they do not yet capture the person’s intentions.

In the next five years, we will see this accuracy greatly improve: eventually AI and ML will know you better than you know yourself. As Yuval Noah Harari states, “the most important thing for humans is to know ourselves. For the first time in history, we will be in competition with who knows us better. It is likely to be the organisations with personalisation engines who know us better than we know ourselves.”

As such, the introduction of an ethical framework for AI and related toolkits will be essential to help organizations execute an impactful customer experience well and without bias. The Apple card, backed by Goldman Sachs, is one example of what biased AI applications can do. It gave, on average, ten times the credit limit to men than to women with the same financial status. The ethical AI framework is designed to mitigate against such biases. We help leaders set up ethical approaches to emerging technologies.

Currently, consumer rights sit within the organisation’s domain. So, for example, a customer who is interacting with a bank will find recourse in laws and regulations relating to the banking sector, rather than in the technology they use to interact with the bank.

Under an Ethical AI framework, consumers should have the right to contest if they think that an AI has made a biased, discriminatory, unfair, or wrong decision about them. The framework can also inform them that they have this right.

The consumer can request that the organisation be transparent with the AI, algorithms, and data used by the AI. The organisation must be able to explain how the AI came up with the decision, and then the organisation (and potentially the AI vendor) will be held accountable for the decision.

You can imagine, en masse, this would pose significant problems for any organisation. For example, there is an AI app called Babylon that registers people’s vital signs and notifies medical help if the person is about to have a heart attack. The app registered symptoms for men as being “heart attack” and for women of the same age as “panic attack.” This was a result of the AI being trained on historical data that had bias within it. These types of oversights and errors can effectively hurt or kill people. It is really not ok.

While there is currently little legal recourse for the consumer when a biased decision is made about them, I believe it is extremely important that consumers contest all decisions where they think an AI has made an unfair decision. The more consumers who do this, the more likely we are to see legislation put in place and have organisations held accountable.

An Ethical AI framework will help mitigate risk for consumers and for businesses. In addition to this, the framework should also provide benefits to the consumer including provision of tailored offers, faster service, more accurate information, and better resolution of problems. Some examples of this include Humana and Cogito, two AI customer service apps, which claim to improve customer experience by about 28 percent each.

Within the next two years, AI is expected to conduct approximately 30 percent of all customer interactions, so it is essential that Ethical AI frameworks are put in place soon.

Start your career in IT today with these online self-paced courses.