Andrew Ng, chief scientist at Baidu Research and professor at Stanford University said, “AI is the new electricity”. Why? Because electricity transformed almost everything 100 years ago and today, AI will quite likely do the same for most, if not all industries.
Artificial intelligence, or AI in short form is entering our lives from every angle, be it healthcare or shopping, and anything in between.
Although we come across the term lots, if you had to explain to someone the differences between AI, machine learning and deep learning you might be somewhat lost for words. So, here’s one of many definitions. AI is “A set of advanced technologies i.e. algorithms, that enable machines to grasp and understand human intelligence, learn to imitate it, and act accordingly.”
Weak or strong?
AI is generally divided into two forms:
• Weak AI, or narrow AI. This type of AI is non-sentient, meaning it is not able to perceive or feel things. Weak AI includes narrowly defined tasks like image recognition, or route planning.
• Strong AI, also known as Artificial General Intelligence. Strong AI is sentient and capable of performing the same intellectual tasks as a human being – and more!
Machine learning is comparable to a person learning from their mistakes. The software and underlying algorithms evolve based on performance and uses that data to improve, beyond that what was programmed by a human.
Deep learning is a sub section of machine learning that tries to imitate the human brain. The algorithms are inspired by the structure and function of the brain called artificial neural networks. These neural networks can learn and make intelligent decisions on their own.
Earlier this year, the Express reported scientists had built a self-aware robot using deep learning to create the self-awareness. Who knows how long it will be before strong AI will become commonplace?
AI development seems to be all around us suddenly. What has triggered this sudden wave of development? From a technical perspective, there are three things that drive the current AI push. The first is that we have much greater computing power than we had even ten years ago. We have the Cloud to store things and Edge computing to expand data storage capability and offer real-time computing.
The continuous rise of big data is another driver. Every device we used generates data which is captured and used to inform, market and improve things. We cannot deal with this level of data without machines taking over the time-consuming data-crunching part.
Many companies now use algorithms to predict behaviour and outcomes. They now have access to more data than before to improve existing models and develop more complicated versions on the back of it. Algorithms need data, lots of it. Some companies use product launches just to get data and monetise that in future launches. Loyalty programs do a similar thing. Data is money.
So, what’s slowing AI adoption down?
At the moment, there is (still) a lack of data to feed into algorithms. Also, at the moment, there simply aren’t enough people skilled in AI to develop customised AI solutions, or even in AI in general.
The consumer angle
How do consumers feel about artificial intelligence? Well, for one, they don’t realise they’re using it.
33% of consumers think they’re already using AI platforms, whereas in fact 77% are already using an AI powered device1. That’s quite a difference.
41% of consumers believe AI will improve their lives in some way2, although they have concerns about their privacy, data security and the potential lack of control of AI technology.
The good, the bad and the ugly
From a business perspective, AI technology is thought to increase business productivity by 40%3. But where is AI being used already? Where is it making a real difference to consumers’ lives? And where is AI going ‘to the dark side’?
Let’s start with the good, there are many great examples. Healthcare is an industry where AI already has a massive impact. In hospitals, AI can help increase the speed of management, diagnosis, research and monitoring, not to mention the speed of discovery and testing of new drugs.
Healthcare has become much more complex and costly, and patients are encouraged to become actively involved in their own care. The Internet of Things has led to wearables which, for example, people with diabetes can wear to monitor their glucose levels (e.g. FreeStyle Libra, Dexcom G6).
Some standard operations are done by robots as well as routine tasks like scans and data entry.
There are also developments on the care side. For example, a companion chatbot has been developed for people with Alzheimer’s, called Endurance. The chat bot has casual conversations with patients and can identify variances in what the patient said previously and highlight areas of memory loss. Doctors and family members can access the conversations also, because Endurance is cloud-based, which helps monitoring.
Furthermore, AI based conversational software can alleviate the data burden for hospitals and GPs. Manually organising electronic health records (EHR) is not conducive to running an effective practice or trust. As well as reduce their administrative burden, GPs and doctors want to increase face-time with their patients. Removing the extensive note-taking can make a big difference.
Our client 3M Health Care has recently bought M*Modal, a healthcare technology provider of cloud-based, conversational AI-powered systems. This records doctor/patient conversations and adds pertinent information to patient medical records, saving time and increasing record accuracy. Doctors piloting its service have reduced time spent on medical notes by 70%.
A great strategic move to stave off competition from the likes of Amazon and Google!
Apart from healthcare, AI is present in many other industries, protecting us and making us more efficient.
In communications, automatic spam filters help us to receive the items we really want to receive and get real-time translations, even though the latter can still be a bit clunky!
In sport, computer vision referees can catch more fouls and manage the game better. Smart ticketing helps fight fraud.
AI can help to eliminate fake news in the media industry, remove journalistic bias (although strictly speaking journalists shouldn’t be biased but report the facts) and even automate journalism, writing the content!
Self-driving vehicles have been in the news prominently over the last few months. Again, it is AI that makes this possible. Other examples in the transport industry are ride-sharing apps and traffic analysis which helps reduce travel times.
In education, AI can keep students honest, with plagiarism checkers. The burden (and potential bias) of marking can be removed through automated grading, and virtual teachers may replace real ones. Not sure how I feel about that, some of my best school memories are of teachers, their love for their subject and their idiosyncrasies.
No doubt there are many other great examples of AI, but let’s see what the British consumer says about AI, and the potential downside of it.
Consumers are certainly a little wary of an AI future. Who can blame them? AI could put 30%4 of existing jobs at risk by 2030 as we move from physical to cognitive labour.
A 2018 study by Accenture titled “Reworking the Revolution” makes a completely different claim: AI in combination with human collaboration may result in a 10% employment increase worldwide by 2020! It’s safe to say we simply don’t know, just that there will be a transformation, which has already started and there’s very little we can do to stop it.
Are consumers ready for the age of the machine? We still prefer to deal with humans over automated services for many purchases, like:
- buying a product or service for the first time (77%)
- chasing an order (73%)
- querying a bill (85%)
- changing account details (62%)
- making a complaint (84%)
That said, we don’t mind so much if we chose the machine interaction ourselves, for example with our smart home speaker, rather than the forced interaction with the chatbot.
For marketers, this means we have to rethink the customer journey to make the machine interaction a credible alternative. Who know, at one point consumers may no longer be able to detect the difference! Alternatively, we can delay the move to the chatbot if it creates friction with the brand promise and doesn’t add significant value to the experience. One thing is for sure, we have to take consumers with us on the journey.
We are in the digital age and spend a lot of time on devices connecting with each other and the world in general using social media. AI is used within social media platforms to process the vast amounts of data. Unfortunately, it can also be used for social manipulation. The most high-profile example of this is Cambridge Analytica and others associated with them. They used the data from 50 million Facebook users to attempt to influence the 2016 U.S. presidential election, as well as the Brexit referendum. I suppose we’ll never find out whether they did. Although AI is not necessarily the ‘bad guy’, it is used to further certain people’s or companies’ less than ethical goals.
Other things to consider are:
- Potential inequality – who makes the money from AI?
- Humanity – will AI support or exacerbate addiction or certain behaviours?
- Reliability – poorly designed autonomous systems can lead to ‘artificial stupidity’
- Transparency and privacy – is data protection a misnomer?
The downright ugly
The media are quick to report scare stories, which are giving AI a bit of an image problem. That’s because as well as being helpful, machines can be programmed to do something dangerous or bad (Ex-machina, anyone?). Autonomous weapons are already here and being used. An individual or country will little regard for life may wish to develop them and program them to kill innocent people.
Another, particularly worrying example of AI is happening in China right now. It’s called the Social Credit surveillance system, it is behaviour-driven and monitored through a country-wide network of surveillance cameras. It gives citizens a score based on e.g. what they post online, whether they buy Chinese products, and whether they flout the rules.
Citizens with an adequate score are considered ‘trustworthy’ and rewarded with better interest rates from banks, discounts on utility bills and travel permits.
However, those with a low score can – reportedly – be stopped from buying a house, using the internet and travel. Nearly 11 million Chinese can no longer fly and four million are barred from trains!
The program was in test-phase but is expanding nationwide (using a network of 600m cameras). By 2020, all citizens (1.4bn) will receive a personal score. The words “invasion of privacy”, “grading”, and “discrimination” spring to mind. Remember, this would never have been possible without AI.
So although AI is improving our ability to look after ourselves and others, keeping us honest and making us more efficient, we may hand in a chunk of our personal freedom in the process. What is your opinion? Follow our blog and comment.
Written by Inge
2 Source: Strategy Analytics
3 Source: Accenture
4 Source: PwC, jobs in the UK
All images courtesy of iStock.com