AI is assisting people in being more efficient and remove the need for banal repetitive tasks, giving us more of our most valuable currency, time.
Imagine you own a brewery that creates bold new eccentric craft beers. A lot of time will be spent researching and inventing new and creative flavours, bottle artwork, and names. Imagine you have created the craft beer that is going to take over London, yet you cannot think of an apt name for such a maverick beverage.
That is where AI and specifically, machine learning, revolutionises the process. Having been fed a data set on previous drink titles, AI uses this to create its own rules on constructing a new name. However, you are not restricted to building the new with the foundations of the past; neural networks can be fed additional information to produce an extensive catalogue of relevant new potential names.
Using a neural network to run machine learning algorithms, AI begins its task much like a child begins their first day at school. Young, naïve, and inexperienced. It has a bold first attempt at the exam, only to fall short of our complex human expectations. It fails fast but learns quick. With practice, revision, and tweaks, AI can learn from its mistakes and consistently improve its effort, creating something that appeals to our cultural and linguistic ideas.
By doing this, we give ourselves a relevant selection of source material to draw inspiration from, which is similar to how we often drive creativity. Contemporary singers are often influenced by previous artists, writers often incorporate elements or concepts of previous authors.
Keeping Humans in the Loop
AI is going to help humans, not replace them. AI can produce material, but cannot critique it like a human, nor fully understand cultural or linguistic nuances that may render it unsuitable. This is where the human in the loop is still the salient piece of the puzzle. A person can distinguish whether language contains innuendo or will be perceived as offensive. A person can judge and combine a multitude of different components from each result to create something better.
“We are on the verge of the data revolution with artificial intelligence, machine learning and data science making their mark across many different business and industries. We see it every day — businesses need to understand that technologies and platforms are great, but there is still a need for humans to be in the loop.” — Kim Nilsson, CEO, Pivigo
There are sometimes fears that AI is going to replace jobs, and this is true, it will. However, it is going to create much more roles than it takes. A study by Gartner predicts that “In 2020, AI will become a positive net job motivator, creating 2.3M jobs while eliminating only 1.8M jobs.” In addition to this, AI is transforming existing roles, helping us to work smarter, not harder.
[Source: Forbes]
Is AI Great at Everything?
In medicine, AI is being used to comb through gigantic data sets of scans to identify subtle anomalies that lead to disease diagnosis. This eliminates human error and saves time, ultimately, saving lives. However, AI should leave the naming of these conditions to the humans.
Research scientist, Janelle Shane, runs the hilarious AI blog, AI Weirdness. She uses Machine Learning to find everything from new pub names to Harry Potter spells. Whilst adept at diagnosing diseases, it is evident that AI has not yet learnt how to label them. The following ailments are the conception of a data fed neural network:
Lower Right Abdomen Degeneration Disease
Cancer of the Diabetes
RIP Syndrome
Loon Eye
Cancer of the CancerT
eenager’s Diarey
Super Syndrome
Spleen Sink
Cysts of the Biles
[Source: AI Weirdness]
Now, most of these sound more plausible as B-list horror movie titles rather than genuine disorders. RIP Syndrome sounds like a way of discussing death in front of children without them realising the nature of the conversation. Perhaps it is because many illnesses are named after the discoverer or often use names that contain terms completely unrelated to the symptoms that make it so difficult to replicate conventional sounding titles.
It is important to recognise the areas that AI can create viable services that improve or revolutionise its predecessor. Recently, it was revealed that new software can lip read with a much smaller percentage of error than humans can, making a real difference to the hard of hearing.
[Source: Verdict]
We are still a while away from the singularity, but machine learning means AI is constantly improving, allowing us more time to focus on being human. AI may still be at school right now, but we have set it a lot of homework.