Reimagining creativity and AI to boost enterprise adoption
AI has yet to reach the point of creativity but continues to advance, while assisting humans in the production of their own creative works and improvement of their organizations.
An AI algorithm capable of thought and creation has the potential to enhance applications and unlock better analysis with less oversight for organizations. However, it still remains out of reach. Until then, AI has an important role to play in augmenting human creativity.
Since the inception of artificial intelligence, researchers have had a goal to create a machine capable of matching or surpassing a human's skills of reasoning and expression. Advancing AI past self-training to computational creativity will require going beyond data augmentation into original thought.
AI, machine learning and creativity
Currently, machine learning specializes in limited data creativity, with algorithms that can train on historical data and allow organizations to make better-informed decisions with analytics. These algorithms use training data sets to "predict" future outcomes and generate new data.
"There are dozens of examples in which different algorithms that, given the observation of real data, are capable of generating very plausible fictitious data, which is almost indistinguishable from real data," Haldo Sponton, vice president of technology and head of AI development at digital consultant firm Globant.
Algorithms can create data, but only when prompted to and only from something that has already been created -- current algorithms can only mirror training data. This falls short of the insular creativity the technology hoped to reach.
To Sponton, creativity is as universal as it is individual. Each being has the ability to be creative, but each individual has a unique approach to creation. Creativity is that ability to use imagination or have original ideas, as well as the ability to create. It is a fundamental feature of human intelligence, and AI cannot ignore it as a step to further advancement.
As AI processes more information, or takes on more intricate tasks, it can evolve and learn to make better decisions. What would make an AI creative is more than just training algorithms and learning outputs, but building from scratch and creating something new, unrelated to existing data.
"This evolution is really valuable, but true creativity has yet to be achieved," said Jess Kennedy, co-founder of Beeline, a FinTech company based in Providence, R.I.
The enterprise potential of creativity
The potential of a creative machine capable of both learning and the ability to create on its own has tremendous potential in marketplace as well as enterprise settings.
A creative algorithm would be able to create data and discover trends without prompting and without supervision. This would mean less maintenance for an organization's data science team and lead to even greater insights, as they wouldn't have to be modeled on existing correlations.
Haldo Spontonvice president of technology and head of AI development, Globant
Overall, a creative AI would have the ability to find the best way to approach most any problem presented to it by an organization. Anything from hunting for anomalies in data sets to prevent fraud to making conversations with virtual assistants feel more natural.
"Tools based on AI algorithms will generate new creative processes, new ways of creating and thinking, new horizons to explore," Sponton said.
At the moment, artificial intelligence has not reached that level of advancement, and the enterprise applications of true creativity are out of reach. Apart from the difficulty of developing an AI capable of creativity, proving that it has had an original idea and is an added level of advancement.
There are some applications of creativity among existing AI technologies. Neural networks are at the point where they can identify tasks in the creative process. Supervised and unsupervised learning can find meaningful connections and patterns within an organization's data set. These systems and approaches have already proven their capabilities in the enterprise, from recommendations for users online to advanced analytics for business intelligence and analytics vendors.
Where we are now: Augmenting human creativity
The combination of creativity and AI has reached an impressive level, but the way we look at it may be hindering enterprise applications. Instead of focusing on developing an AI that can stand alone and be considered creative, experts note that AI is already successfully helping to further human creativity.
"AI has been used to create things like art and music, but it has been based on existing information and data provided to the AI interface in order to do so," Kennedy said.
This allows for the creation of traditionally creative materials by AI but falls short of that ultimate goal of a creative AI. This does, however, allow for a uniquely nonhuman approach to the creation of artistic works.
"Artists around the world are already adopting this technology for musical composition, for the creation of plastic works and even choreographies or sculptures (just appreciate the work of choreographer Wayne McGregor or plastic artist Sarah Meyohas)," Sponton said.
Adding another layer into the field of creative arts opens up new opportunities for expression and beauty for those working in the field. Instead of taking the human aspect out of this field, this augmentation role for AI finds a balance between creative AI and solely human creations.
"The truth is that these algorithms generate new data, such as images or music, which can be considered a result of the imitation of the human creative process," Sponton said.
AI is not at the stage where it can stand on its own and create, but for now, it serves a valuable role of creating data, analyzing processes and augmenting the creation process. When the time comes for an AI to take the next step, however, we may even have to redefine creativity.
Strategies to successfully deploy AI in the enterprise