An interview with Andras Kenez: Artificial intelligence is here
Andras Kenez is a marketing consultant and lecturer at Staffordshire University’s business school. What follows is an AI interview with Andras for our book, The Conversation Engine.
Match.com recently launched Lara, a virtual dating assistant whose role is to give singletons a helping hand in finding a partner. The (artificially) intelligent Lara talks to members to build up an understanding about their interests and profile.
Analysing this information, she can help to find the perfect match. Lara uses natural language and speech recognition, so users feel like they are talking to a real person.
I took a similar approach when I implemented a chatbot for my students at Staffordshire University’s business school. This chatbot answered frequently asked questions about the module such as:
- What is the deadline?
- How do I submit the assignment?
- What is the word count?
This meant that my students no longer needed to open the university pages and read long documents to find the answers to simple questions.
Artificial intelligence (AI) is no longer the future
Artificially intelligent computer systems are able to perform tasks that normally require human intelligence, such as visual perception, speech recognition and decision-making. Ultimately, it is the technology’s ability to learn over time that sets AI apart from basic algorithms.
We regularly hear AI discussed as a future concept, or something yet to really take off. However, AI is already here, and you’re interacting with it daily.
In 2016, Facebook launched an open version of Messenger that invited developers to create AI chatbots that interact with Facebook users. This service means any business can now engage their customers instantly — even if the responses aren’t as personal as real human interaction.
The technology giants are the best examples of companies using AI and the data gained through visitor interactions. Facebook already has data on you from your profile, page likes and messages. The adverts you see on Facebook have been selected just for you, based on your interests and previous patterns — it’s no coincidence that you see products you were looking at a few days ago pop up in your timeline.
Equally, if you search something on Google, the results that you see aren’t random. They are based on your previous history and behavioural patterns; unless you choose to go incognito on Chrome.
Similarly, Amazon displays a lot of offers based on what it thinks you may like. If you buy three items, AI can learn what other users who also bought these items subsequently purchased, and suggests these to you too.
Chatbots using AI may have even sold you something without you noticing it. Information on how you respond to these chatbots can be used to gain customer insight, and store information about you to personalise services and give bespoke offers in the future. Combining people’s predictability with clever planning means chatbots are able to answer many customers’ queries.
The compromise of supervision
AI does have its limitations at present. How an AI bot collects its data has huge implications on what it can learn. The basic bots only do what they are told, they are not able to learn and therefore the knowledge of the human programmer is a limiting factor in the bot’s development.
Other bots, on the other hand, use big data to learn, meaning the whole online community is a part of developing the knowledge. Unfortunately, this large-scale approach does mean that there is much less control over what the bot learns.
Some high-profile AI scandals have made the headlines recently, warning that AI robots are becoming sexist and racist. As written in the Telegraph, Maxine Mackintosh, a leading expert in health data, believes that this problem was mainly the fault of skewed data being used by bot platforms, saying, “These big data are really a social mirror [sic] — they reflect the biases and inequalities we have in society.”
This is the sad realisation that the internet itself is sexist and racist, not the AI bots.
Humans have an important role in supervising AI throughout the learning process. It’s not just about preventing racist and sexist scandals, but about keeping the bot in line with business goals.
If a self-learning AI bot is completely unsupervised and free to learn from a wide community, then you may find it convincing your customers to buy cheaper items or even to choose a different supplier entirely. This would defeat the purpose of investing in such technology, and you wouldn’t see a return on investment.
Ultimately, you need an artificially intelligent sales agent for your business, and not an artificially intelligent impartial third-party spokesperson.
Out and about
There are many other sources of data to add to the AI equation. For example, geolocation signalling plays a huge role in personalisation of service. Knowing where and who you are allows businesses to send you real-time, tailored offers. So, if you love coffee and are around the corner from a coffee house, then expect to be sent an offer to entice you to get your favourite iced frappé.
AI is also interested in how people look. Tesco was the first to announce its plans to tailor adverts using the facial recognition tool OptimEyes. BP, Shell and Esso soon followed suit.
These hi-tech screens scan customers’ faces in petrol stations so that advertisements can be tailored to suit them. The technology also adjusts adverts depending on the time and date, to give a totally bespoke advertising experience.
Here and now
While the media often talks about the ‘future’ of AI, it’s important to see that this technology is already here. It is the big companies who are already taking advantage of this powerful tool, leaving smaller companies at the starting blocks.
However, in the future, we will start to see AI technology becoming more widespread for all, and not just a privilege for the largest conglomerates.
Novelty is what drives AI chatbots at present: it’s new. However, one day AI bots will become the norm, and perhaps even stale. This technology needs to continually advance, and the way we use it needs to evolve to stop it from becoming irrelevant. We are starting to see this already, with some exciting opportunities appearing in the entertainment sector.
As part of marketing Disney’s Zootopia movie in 2016, an Officer Judy Hopps chatbot was launched to engage the audience. This allowed fans to chat one-to-one with Judy, who delivered in-depth and immersive storytelling experiences. Buying this time with consumers meant they were more likely to go to the film, tell their friends and buy merchandise.
This demonstrates the huge potential in AI partnering with entertainment. It’s not all about on-the-spot deals, but rather the customer experience as a whole.
Whatever way you look at it, AI is here, and here to stay.
NB: This AI interview is only an extract from The Conversation Engine. To read the full book, download a free copy here: https://www.whoson.com/the-conversation-engine/