Click here to print this page

Planning Retirement Online



Everything you always wanted to know about Artificial Intelligence but were afraid to ask

By Tony Clack,
retired founder of LaterLife

This article is based on topics and information
on the website

December 2017

Why do you need to know about AI (Artificial Intelligence)?

AI will impact every aspect of our lives and at a pace never seen before. AI and the associated technology of robotics are being talked about as key components of the 4th industrial revolution because of the level of change and the level of impact they will have.

In 1951 the first operational run of a commercial computer took place. It was produced by Lyons and called LEO (Lyons Electronic Office). Today as we all know computers are totally pervasive both in the workplace and in our personal lives. The progress of AI is predicted to be much much faster. There are likely to be society changing impacts within the next 20 years, and the process has already started.

It has huge potential to change our lives for the better, inherently being able to improve quality while reducing costs. In the opinion of many it can be the answer to most of our big problems from the Health Service and Care Homes to Transport. But like all advances there can be downsides and unintended consequences and we need to plan to avoid these right now because they are potentially very big. At a personal level we need to understand AI systems and how they can help us and we also need to be able to trust them, especially when they are dealing with fundamental aspects of our lives.

So, what is AI?

Forget the idea of complete replacement for human beings for now. There are many areas of Artificial Intelligence and these are all at differing stages of development and cover a spectrum which includes:

  • Ability to adapt behaviour
    (systems which can change the way they operate based on experience)

  • Ability to analyse past examples and learn
    (deep learning to construct and apply new strategies)

  • Ability to make sense of and interact with our world
    (Speech and Object recognition, Touch etc.)

  • Ability to apply learning from one environment to others
    (and therefore tackle any problem)

  • Ability to understand and reproduce human behaviours
    (using the above coupled with robotics)
AI solutions may use some or all of these and range from software-only internet solutions to autonomous robot solutions and augmented human intelligence implants.

This is not the place to try and explain all of these but we’ve tried to provide a simple layman’s explanation of the first of these as a supplement to this article below.

How will it impact us?

The answer is in every corner of our lives and in ways, some of which, we currently can’t predict. Jobs for example will be transformed or replaced.

The impact on the transport system e.g. self-driving cars, is reasonably easily imagined, similarly autonomous ships and trains. As well as the impact on us as users, consider the impact on associated transport jobs.

Medical diagnosis is another obvious area where it is already being shown that systems can achieve a higher diagnosis accuracy than doctors. Even if we use these systems in a triage way initially, where the doctor still reviews the suggested diagnosis and recommended treatment before actioning it, think of the money that could be saved in the Health service at the same time as improving the quality of diagnosis.

In Care Homes, robotics can potentially be applied to many of the standard caring activities e.g. toileting, bathing, turning in bed, providing refreshments, reminding when to take medicines, fetching cups of tea, cleaning, bed making etc. The robots would also be on call 24x7, would not get annoyed, would ensure the same high standard of care, wouldn’t be rough in handling you, would talk to you in a nice soothing voice etc. Potentially this could both reduce care home costs and free-up staff to talk more to the elderly in their care, improving the human interaction aspect. Personally, I’d prefer a robot bathing and toileting me rather than a human. Would you? You can answer this and a mix of light-hearted and serious questions below.

The potential impact is endless and will spawn lots of entrepreneurial activity. It also comes with the silver bullet of reducing costs and improving quality and effectiveness at the same time.

The big question is what is it going to do to jobs? There is an argument that says because we’ve seen these revolutions before, it will be the same, many jobs will change and be augmented by AI, others will cease to exist because they will be automated, or more likely replaced with new improved ways of doing things, while lots of new jobs will be created.

However, there is a counter argument that this time there is a fundamental difference: AI in theory can replace any job, some jobs will be more easily replaced in the early days e.g. the medical diagnosis example. However, as the technology develops it will reach everywhere, so it is totally pervasive across jobs, and any new jobs created will themselves be open to being done by AI.

This technology puts huge commercial power in the hands of its owners. If we do nothing to change, all the money and power is likely to go to a very few individuals and multi-national corporations with consequent impacts on society.

In my personal view we need to embrace AI technology because it can have huge benefits for society and address many of our current big problems. At the same time, we need to be looking ahead with urgency to understand the likely impact on jobs and society and to decide how we ensure the benefits of adoption translate into benefits for all of us.

Questions for you to answer

  • Could you tell the difference between a human in a contact centre answering questions by email and an AI application?

  • Would you prefer to be bathed and toileted by an AI robot or a human in your old age?

  • Could you befriend an AI robot if it showed empathy?

  • Could an AI application and/or robot do part or all of your job, or the one you used to do?

  • Would you be happy for AI to perform the triage function in your local doctors?

  • Could most of Government be replaced by AI?

  • Would political robots have to be barred from answering questions?


To give your reactions to these just visit the
Get it off your chest’ discussion on Probably42.

You can become a member of for free and add your own comments and ideas on a range of interesting topics.

Probably42 is the first initiative of ‘The Knowledge, Skills and Experience Foundation’ which is a non-profit making Social Enterprise organisation set up by Tony Clack and Dave Sinclair, retired founders of LaterLife.  


Supplement on understanding AI

We’ve taken just the first aspect of AI listed above and tried to provide a simple layman’s explanation of the first of these:

Ability to adapt behaviour

Traditionally computer systems follow a set of rules within a computer program known as an algorithm, and results or behaviour are totally predictable given a specific set of inputs. Adaptive systems, however, can change their own algorithms based on experience e.g. by changing the weighting of different factors used in the decision making. The aim is to make them more effective in a particular well-defined environment. The strategy used for making changes is intended to gradually improve the way the program works but isn’t guaranteed to do so.

Let me take an example to illustrate this (note this is not scientifically accurate but illustrates the point):

If we take Voice Recognition (essentially pattern recognition), then a lot of research has been done to enable creation of algorithms which can recognise many different voices and different accents but aren’t perfect and these are still developing.

However, some systems using voice recognition might have been programmed to capture what recognition tasks they have performed, whether it resulted in success or failure etc. and do an automatic analysis across many instances of attempted recognition to identify patterns of success and failure. Then to try out slight changes to their algorithms which may improve the chances of correct recognition and then use these modified algorithms in future.

Further experience can then measure, in the same way, whether there was in fact an improvement in recognition and make further adjustments accordingly. Thus, the operation of the system adapts itself over time in ways that the programmer could not predict. On occasions it may go backwards because what it tries doesn’t work but over time it should keep improving.


Read more articles here.

Back to LaterLife Interest Index

Bookmark This Share on Facebook Receive more like this

Latest Articles:

Gene therapy - the future of our health

scientist in lab

Gene therapy is hugely exciting. Whether it will fulfil its promise and in future years produce terrific treatments for many health problems we don’t know but at the moment, although still in its early stages, the results are very encouraging.


Art Masterclass
Spring Flowers

Step by step guide to creating a beautiful spring flowers landscape.


Listen free to uplifting joyous birdsong

Turtle dove

It is great news that we can now all enjoy natural birdsong at any time of day and night thanks to a new birdsong audio that the RSPB have put together.


Win great prizes in our current Competitions

Silver Travel

Click here to visit the competitions page.

Article Archive

The LaterLife Article Archive provides a comprehensive list of links, to all the current regular article series' as well as quick links to older articles.
Back to LaterLife Today

Visit our Pre-retirement Courses section here on laterlife or our dedicated Retirement Courses site


Advertise on