Wiser Technology Advice Blog
- ARTIFICIAL INTELLIGENCE EXISTS ONLY IN MINDS OF MARKETERS
Artificial intelligence exists only in the minds of marketers
Artificial intelligence (AI) really exists only in the minds of marketers. Machines are smart, but not really intelligent. Artificial intelligence is just a marketing buzz-word, whilst we have very smart machines and systems there is nothing (yet) that is as intelligent as a human.
What is artificial intelligence, really?
Firstly, what is Artificial Intelligence? It could be argued that there is no such thing in reality, this is a very broad term that's been made up by marketers, like 'Cloud Computing' is just a marketing term. Most artificial intelligence systems are really just machine learning algorithms. Some systems are more advanced, using what's termed deep learning or neural networks.
But all of today's artificial intelligence systems have a very narrow scope. For example, my robot vacuum cleaner cannot talk with my coffee machine or internet television. True artificial intelligence will only come when the internet of things is able to connect all the systems and we have smart systems that combine all the devices. So don't panic, The Terminator singularity still is many years away!
An early example of artificial intelligence used hand writing recognition to automatically sort the post into post code regions. Machine learning algorithms are trained by providing them lots of examples, such as the various ways numbers can be written.
The algorithm must then be tested to check if for example it can spot all the number sevens in the image. Any exceptions show where there were gaps in the training data which must be fixed before the algorithm can be relied upon.
More advanced artificial intelligence uses deep learning, which is being used for image recognition.
Image recognition systems make it more obvious why we use training databases for artificial intelligence, rather than the traditional methods of software coding.You simply cannot program what a dog looks like, as the wide variety of what a dog looks like cannot be described in words. Artificial intelligence system training is done by providing lots of flashcard examples. It's like teaching a toddler to recognise dogs, by pointing out all the dogs we see.
Did you know that the Google Recaptcha "I'm not a robot" proof you are sometimes asked to do when putting in an online order, where you click pictures with images of cars, street signs or shop fronts, is you training the Google artificial intelligence image recognition systems?
Image recognition of cats poses similar challenged, as cats have very similar attributes to dogs. So again, we provide lots of examples of images with cats in the training database.
Larger training databases give better accuracy, but a word of warning, beware of unconscious biases in your training database. There have been famous (infamous) examples of voice recognition systems not understanding people speaking with a non-American accent and image recognition systems not being able to interpret images of people with dark skin.
A deep learning algorithm functions as a black box, and it cannot explain to a human how it got to its answer. An analogy of this is when you're pondering a difficult problem all day and cannot find a logical answer, but then wake up with an insight into the problem after getting a good night's sleep. The deep recesses of your brain have been working hard for you while your frontal cortex was resting, but you won't know exactly how you came up with the answer.
Artificial intelligence is however not perfect and we need humans to provide judgement on the suggested answer. The example of recognising a dog or cat can provide a fairly certain outcome, but it's not going to have 100% accuracy.
What is available now?
There are a lot of appliances and systems that are marketed as having 'artificial intelligence', but in reality these are very limited in their scope and most have just a little bit of smarts built into their operation. For example, Samsung are marketing one of their latest washing machines as having 'AI personal settings', which in reality is a little bit of machine learning to remember your favourite wash cycles. Whilst that could be a very handy feature, it's not really artificial intelligence, is it!
The least sexy but common and very valuable use of artificial intelligence for business is simple machine learning algorithms. The training database is provided with examples of data points we expect to match, then the machine learning algorithm builds itself a set of rules over time, requiring less human input to verify data matches as it learns.
The very real business benefits from simple machine learning algorithms allow you to
- Cut loose the parts of the business which are running at a loss.
- Identify trends with many small issues that are creating unrecognised large and wide-spread problems.
- Get a better understanding of the demographics of your customers, enabling personalised, targeted marketing.
- Manage your spending better by understanding what is purchased, how often and from whom, enabling better supplier deals to be negotiated.
Which system is right for you?
There's so much that can be done with 'artificial intelligence' systems that it's easy to get carried away. But my advice is to be sure that you will get a return on investment from the introduction of new technology and systems. Before committing to spend money with a technology provider you need a robust business case, with cost benefit analyses over multiple years that compare total cost of ownership to the potential for savings. The world of information technology can be complex and overwhelming. Procurement of information technology is a journey that needs wise advice to be navigated safely.
Wiser Technology Advice is here to help! I do the research on your behalf, so that you're not overwhelmed with choices and not badgered by technology companies you meet at conferences and trade-shows. One great way to introduce change whilst managing expectations is through business process mapping workshops, just one of the great consulting services I offer. Or, if you're ready to invest in technology but don't know where to start, I can provide you with independent, trusted advice, taking the risk out of technology investment.
But wait, there are risks to consider…
Deep learning is where artificial intelligence starts to get a little scary. Companies such as Google, Facebook and Tesla don't understand what their artificial intelligence algorithms are doing, because humans can't understand deep learning artificial intelligence algorithms.
And these deep learning algorithms are only as good as the data they're given in training. Human bias can inadvertently be introduced. Many algorithms which are trained to do a human job, such as recruitment, credit scoring, assigning a prison sentence or diagnosing a patient, will have been trained on human generated data, from past occurrences of a person doing the job. If they are trained on data which contains human bias then of course the algorithms will learn it, but furthermore they are likely to amplify it. This is a huge problem, especially if people assume that algorithms are impartial and won't have the same biases as people.
When artificial intelligence evolves to self-awareness, human beings could be in for big trouble. Bill Gates (founder of Microsoft), Elon Musk (Tesla) and Stephen Hawking have all been quoted as being very scared by artificial intelligence.
Imagine a world where a robot has self-awareness and has been given a task to perform. Nothing will stop that robot from completing its task. Humans will not be able to stop the robots. Boston Dynamics have already developed a robotic dog that can't be stopped from opening a door…
Want to know more?
Please get in touch if
you’d like my help to explore how your business can take advantage of smart machines and systems, there's a broad range of solutions and something to suit every business in every industry.
Get in contact today, I’m always happy to meet and have a chat over a coffee.
AI is not just learning our biases; it is amplifying themhttps://medium.com/@laurahelendouglas/ai-is-not-just-learning-our-biases-it-is-amplifying-them-4d0dee75931d
Only a fool would dare interrupt Boston Dynamics' SpotMini from opening a door https://www.theverge.com/tldr/2018/2/20/17033982/boston-dynamics-spotmini-door-opening-video-interrupt-test
Samsung Bubblewash Front Load Smart Washing Machine with AI https://www.samsung.com/au/washers-and-dryers/washing-machines/8kg-smart-ai-front-load-washer-ww85t504dae-ww85t504dae-sa/
Tech giant CEOs willing to set AI standards, ACS Information Age 26 Mar 2021, available at https://ia.acs.org.au/content/ia/article/2021/tech-giant-ceos-willing-to-set-ai-standards.html
What the Rise of Sentient Robots Will Mean for Human Beings https://www.nbcnews.com/mach/tech/what-rise-sentient-robots-will-mean-human-beings-ncna773146