With the increase in awareness, conversation… and fear of AI, I want to discuss the pros and cons of AI, now and in the future.
There are four factors driving Artificial Intelligence today:
The AI Winter is over, funding for AI is not solely coming from academia any more, venture capitalists, investment banks and billion dollar corporations are investing heavily in AI.
Moore’s law states that the number of transistors in a dense integrated circuit doubles about every two years. This is creating exponential technical growth all around us. To put this into context, by 2021, the processor on a £1,000 laptop will have the computational power of the human brain. By 2045 it will have the power of the human race.
As a species we are creating an unprecedented amount of data, we have created more data in the last two years than in the entire history of humankind.
Every minute we are sending 15 million messages and emails, we are uploading 300 hours of video to youtube and 240 Billion (yes billion) heartbeats to the cloud.
Human beings are embracing AI as it is all around us, in the phones we use, our voice operated assistants like Alexa and Google Assistant, our fridges and the apps that we use.
AI is controlling our social feeds, the partners that are selected for us and the news we consume which is affecting the way we think, feel, vote and fall in love.
As a digital agency that is spending more time focusing on AI, we have implemented it with data, voice (using Alexa and Google Assistant), personalisation, chatbots and computer vision.
There are countless examples of these and more all around us.
Eminent minds around the world are arguing about the use of AI, some think it will cause the end of civilisation, others think it will transcend us to “Homo Deus”.
There are countless organisations discussing the ethical implications of AI and its influence and effect on us.
So, with all of this being said, does AI suck, or does it rock?
Expectations Are Too High
AI is all around us, definitions vary wildly. Some people think that the robots are coming and that AI is capable of sentient thought. As a result of this, expectations of its capability can be overestimated. For example, if you say to Google Assistant "I’m having a heart attack" it says "I hope you feel better soon!" Some people’s expectation is that maybe your phone would turn into a defibrillator! We are not there… yet.
The AI — Human Gap Is Growing Exponentially
With Moore’s law and the incredible amounts of data being uploaded, segmented and analysed, the capabilities of AI are ever increasing, however humans learn in a linear fashion and the gap between AI proficiency and our understanding of it is ever-widening.
Solutions are being created which feel so complex, (take blockchain for example) that it is almost like they have been created for problems that don’t even exist yet.
This growing misunderstanding could create mistrust and uncertainty in AI as it gets more advanced.
In order for AI to make decisions it needs to have knowledge. This knowledge is based on data and it learns based on data in much the same way that a child does. If you tell a child one animal it sees is a dog and another is a cat, that child will be able to tell the difference in the future. In the same way, if you upload 1000 pictures of dogs and 10000 pictures of cats to a machine learning algorithm and categorise each of those images, the AI will learn the differences and when it is shown an image of a cat or dog it will be able to differentiate. This only works if you categorise the images correctly. If you uploaded 1000 images of rabbits and categorised them as cats, the AI would ‘think’ rabbits were cats.
Above is a Google image search result of the term ‘CEO’. So here you can see where potential problems lie. The data that Google already has on CEOs means that the results returned are almost exclusively images of white men in suits. Is this correct? Well, over 90% of Fortune 500 companies have white male CEOs, but that doesn’t mean all CEOs are white men.
So, with this ‘biased’ dataset, if you uploaded 1000 images of people to an AI based on Google’s data, it would think that any people who aren’t white men would likely not be CEOs.
This opens up a huge debate on the bias that we all have as human beings, but if it is human data (videos, images, messages, reviews) that teach AI, what chance does AI have of being truly objective?
Microsoft launched an AI bot on Twitter called Tay in 2016 that learned from its interactions with other Twitter users. It took less than 24 hours for Tay to start spouting anti-semitic and genocidal hate before it was taken down by its creators. This experiment demonstrates, once more, that if you surround a ‘sponge-like’ intelligence with biased data that it can be influenced by it.
Its Stealing Our Jobs
In 2017 The Royal Society of Arts conducted a survey that concluded that four million jobs in the British private sector could be replaced by robots in the next decade.
In 2013, Oxford University conducted a study of how automation could affect jobs in the future, the study found that jobs such as telemarketers, insurance underwriters and library technicians have a 99% chance of being completely automated.
In 2017, Amazon announced that it would be ‘employing’ 50,000 robots.
What does this all mean for the human race and employment? If our jobs are all going to be automated what will we do? If jobs are changing what should we teach our children in schools?
All of these questions create uncertainty and fear in AI and automation.
Its Changing The World (And We Don’t Know How)
AI is affecting our lives in so many ways, algorithms are used to determine our preferences and display what it thinks we want to see, hear and read. AI is controlling our social feeds, the news we read, the films and TV watch, the music we listen to, the potential partners we meet…
Ergo, AI is controlling the way we think, the way we feel and the way we fall in love.
AI is incredibly difficult to regulate, at SXSW in 2018, Elon Musk said that AI is more dangerous than nukes. To build a nuclear missile, you need plutonium, this is rather difficult for most people to get hold of. Any individual with a smartphone and internet connection has access to more computational power than ever before.
Its changing the way we live our lives and none of us really know what this means… we’re scared of the unknown and losing control.
Ok, so enough doom and gloom, let’s now look at what makes AI an incredibly exciting addition to our daily lives and our future.
Expectations are too high
Try telling Apple’s Siri that you are having a heart attack. The first thing it does is display a button that allows you to dial emergency services.
Our high expectations and belief in the incredible capabilities of AI mean that we will always keep trying new things.
This creates a self fulfilling prophecy — if we believe that AI can truly improve our lives, then it will happen.
Gen-Xers, baby boomers and even some millennials may still struggle with voice controlled AI because they are so used to having a controlled navigation or search but just watch a child interacting with Alexa or Google assistant, it is a joy to behold 😀
The AI — Human Gap Is Growing Exponentially
Earlier, I discussed that our linear understanding of technology was being left behind by technology and the gap was widening, that we were creating solutions to problems, as yet, undiscovered. If you think about this, it is like we have created a time machine! If we can truly create solutions to the problems of the future, how can this be a bad thing? When those problems do arise, we will be able to solve them before they cause significant issues.
Even though creating AI in our flawed image could be dangerous, the more attention that bias receives, not just in data but in all walks of life, the more It makes us challenge the way we look at the world.
The more we challenge, the more opportunity for movements like #metoo and #timesup. The more we study ourselves and our perception of each other, the more opportunity for inclusion and diversity in our society.
Its Stealing Our Jobs
Technology has been ‘stealing our jobs’ for centuries.
When the plough was invented farmers were able to use animals to tend their fields, they didn’t sit around and do nothing.
When the printing press was invented, it created the ability to spread messages throughout the world, scribes didn’t just sit on their hands.
When the jet engine was invented, shipbuilders didn’t ‘down-tools’.
Technology has created more jobs that it has stolen, how many of us would be doing what we are doing now if the internet hadn’t been invented? I know I wouldn’t!
Automation of certain jobs can create massive efficiencies, these do not have to mean that companies need to lay people off, they can use the time saved to allow workers to concentrate on more complex tasks, more exciting tasks, tasks that can help change the world for the better!
Its Changing The World (And It’s Very Exciting)
Change doesn’t have to be scary, as Heraclitus famously said, the only constant is change. The internet has created the ability for us all to be connected, we all have the power to work together to solve the world’s problems, once we augment ourselves with AI, we could cure disease, feed the hungry, clothe the poor and educate the masses.
AI can help us to understand ourselves and the world around us, the more we understand, the less we fear. The less we fear the more we can change the world for the better.
I am the founder of a digital agency, Greenwood Campbell. We transform and accelerate brands using technology, for the last year that technology has been AI, data, voice and chatbots. We don’t just work with marketers, IT and CEOs, we work with humans, we spend time understanding their motivation, their emotional connection with a brand and the tasks they want to complete.