Features | News
21 May 2018
9 things everyone should know about AI
Three Phrasee experts give us their insights and reveal the key things that everybody should know about AI.
We live in a world of rapid technological advancement, where space travel technology is making breast cancer screening available at supermarkets1 and where 90% of all the data in the world has been created in the last two years2. Our current output of data is roughly 2.5 quintillion bytes a day3.
People are apprehensive about the scale and pace of this change, with Forbes claiming that rapid technological change just may be the biggest threat to global business, citing a recent CEMS graduate survey4. Confusion is common with terms such as automation, artificial intelligence, machine learning and big data frequently bandied about, without explanation or context.5
But what does it all mean?
To help remove some of the frequently-generated smoke and mirrors from the fast-paced, rapidly growing world of emerging technologies, we asked three Phrasee experts to give us their insights and reveal the key things that everybody should know about AI.
Neil Yager, PhD, Phrasee Chief Scientist and company co-founder
AI isn’t new. People tend to think of AI as something futuristic, or as an emerging technology. However, a lot of the theoretical groundwork was laid decades ago. The exciting advancements we are seeing today are largely a result of better software tools, more powerful hardware, bigger datasets, and increased R&D. There will undoubtedly be breakthroughs in the years to come, but even today AI is a mature field with many practical applications.
While AI can be revolutionary, most of its uses aren’t. It has only been a few years since self-driving cars sounded like a science fiction fantasy, but most uses of AI don’t result in a radical departure from the way things are currently done. Instead, they focus on increasing efficiency or obtaining incremental performance gains.
There is an opportunity cost to not using AI as part of a business. AI is here and it can be used to improve business processes. Companies that don’t adopt this technology will have trouble competing with those that do.
Victoria Peppiatt, Phrasee COO and co-founder
AI isn’t one specific ‘thing’. Instead, it’s the broad term used to refer to a wide range of technologies including machine learning, natural language processing, natural language generation, deep learning, machine vision and others.
At Phrasee we have two forms of AI. Natural Language Generation generates the language that sounds like it’s been written by a human and it can create millions of lines of copy at the same time. Our Deep Learning engine takes the results we feed into it and can learn from thousands of neural networks to discover what’s working and what’s not. So we can optimise subsequent generations based on that knowledge.
Robotics and AI should not be confused. Robots are pieces of hardware, such as the robots that build cars in factories. AI is a software, like the technologies that generate facial recognition on your phone or recommend songs on Spotify or write email subject lines at Phrasee. However, you can combine the two and a robot that makes ‘decisions’ will normally use AI to do this.
Don’t use AI because it’s a buzzword, use it to solve a problem. For example, our clients use Phrasee to increase open rates in their email campaigns and we outperform the email campaigns that have been created by humans 98% of the time. We can say this categorically because we are always measuring against what a human would have done because you only want to be using AI if it’s improving your results.
Stefan Britton, Phrasee CCO and AI product expert
Artificial Intelligence is a tool and not a solution in its own right. Businesses shouldn’t want to ‘buy AI’ because it’s trendy. Instead, they need to think about the business problem they want to solve and consider whether AI is the right solution. Ensure that you can quantify the problem when sourcing a suitable solution, AI or other. Because if you can’t measure it, you can’t improve on it – which means there is no point spending money on it.
AI doesn’t want to rule the world. AI is a form of computing that’s amazing at analysing huge sets of data and automating decision making. AI does not have an ego. It does not want to get promoted. Nor does it want to make you look silly at a dinner party. As humans we transpose our worst traits, expecting AI to “think” like we do. Can AI be used in dangerous ways? Of course, but AI doesn’t have ambitions to rule the world. Rest assured, humans are still the most dangerous part of AI, not the technology itself.
Emotions are still complex for humans, let alone AI. Free will and emotions are so complex that human beings aren’t even close to understanding them. They are such complex issues that, even after decades of studying them, we still have a long way to go before being able to understand and replicate them in AI. We may be able to create algorithms capable of mimicking decision making, but the technical (actual) re-creation of intuition is still way beyond us.
“If the brain were so simple we could understand it, we would be so simple that we couldn’t” – Lyall Watson
So far, so useful. But what does the future hold?
At the end of April this year, Mark O’Connell’s To Be A Machine won the prestigious Wellcome Book Prize, annually awarded to a publication that engages with some aspect of medicine, health or illness in fiction or non-fiction. In it, O’Connell speaks of the moment of “Technological Singularity”, when AI will usher in a “new human dispensation, a merger of people and machines”, allowing us to transcend the “limitations of our biological bodies and brains” and gain power, not only of our fates but also of our mortality.
It’s a long way from using AI to boost audience reach in marketing campaigns, no matter how complex the algorithms. And it’s a rather mind-blowing prospect. But will it happen?
Today we have no way of knowing but we can look to the past to help us anticipate the future. And remind ourselves that 100 years ago personal flying machines and passenger boats pulled by whales had been happily predicted by our predecessors, while in 1900 the authoritative Ladies’ Home Journal expected the future elimination of the letters C, X and Q from the alphabet, seeing them as “unnecessary”.
The reality is that it’s impossible to tell what changes AI will bring. But in the meantime, we can content ourselves with removing some of the smoke and mirrors from the technologies and embracing the benefits to make positive changes, one small step at a time.
1 2018 NHS pilot project sending 16 breast screening vans to supermarkets in three UK regions, with capability of beaming scans directly from car parks to hospital experts using technology developed for space travel
2 Stats from IBM, widely and frequently quoted
3 A number with 18 zeros
4 Survey of graduates from the Global Alliance in Management Education (CEMS) Masters in International Management programme
5 This article is designed to alleviate your confusion