Is AI about to take over? Will the tables soon turn? Will we become the servants of the very machines we have created? Not quite. Such postapocalyptic visions may have caught the public imagination, but the reality still seems altogether different. Self-aware AI is nowhere on the horizon.
Instead, artificial intelligence is overwhelmingly dependent on us, humans, both to feed it information and help make sense of it. This almost makes it sound like a very limited system. It’s important to remember something, though — while it does have practically endless potential, AI is still in its embryonic phase. It has been around for just over 80 years, and in science and engineering, that’s a comparatively short amount of time.
So what exactly is its biggest obstacle to further progress? What is the missing piece of the puzzle? Sensemaking with AI is where the greatest room for improvement lies.
What is sensemaking?
Sensemaking is nothing more than the way in which we, humans, give meaning to our collective experience. It’s innate to us, living creatures. This ongoing retrospective development of plausible scenarios helps us learn and make decisions. As its name implies, it allows us to make sense of our environment and our interactions with it. It tends to happen in three phases:
- Discovering or acquisition of knowledge
- Debriefing or rationally summarizing the experience
- Return to normalcy
The above is only a part of the definition. In order to truly understand it, we need to move a step further. Academic researchers have found out that sensemaking has seven properties.
- Rooted in identity — Sensemaking rests on the foundation of self-conception. No sense of self, no sensemaking.
- Retrospective — It relies on previous experiences. Conclusions are made at a backward glance.
- Ongoing — It needs the past, the present, and the future in order to exist. Such a timeline enables learning.
- Enactive — Actions result in physical or structural change.
- Focused on extracted cues — These starting points serve to generate further ideas which, in turn, become linked and connected into networks of meaning.
- Social — Sensemaking depends on interaction with other individuals. It can also rest on assumed or expected interactions within the group.
- Plausible — Ultimately, sensemaking has to make sense. It needs to at least appear logical and rational.
Can a computer do it?
Sensemaking isn’t the easiest thing to replicate in a machine, right? Much like any other technology, AI isn’t (yet) perfect nor is it even close to replicating such a complex phenomenon. Any artificial intelligence depends on the quantity and quality of data it has access to. Furthermore, AI isn’t exactly blessed with intuition and creativity. To this day, sensemaking is a trait solely reserved for living creatures. It’s exactly this creative spark that forms the biggest obstacle in sensemaking with AI. However, substantial progress has been made.
Examples of sensemaking in AI
There are a few niches in which sensemaking in AI shows its true potential. Natural language processing (NLP) is one such example. In it, algorithms analyze human language (be it English, Mandarin, or Swahili, or any other human language) and derive meaning from it. This is, in turn, very useful in sentiment analysis. AI is able to identify opinions or even detect emotions through written or spoken words.
Computer vision is another area in which AI has achieved great results. Algorithms are able to analyze both static and moving images to identify objects and patterns. AI’s ability to figure out its environment in real time is particularly useful in self-driving cars. Other areas of sensemaking applications include digital diagnostics and augmented reality.
The best ways to leverage machine intelligence
An integral part of AI, machine learning, comes in two forms — supervised and unsupervised. Depending on the form, ML leverages past data or experiences to make new conclusions with or without being explicitly instructed to do so. In the case of sensemaking, the process definitely requires supervision. It’s crucial to understand what AI can and can’t do. After all, machines are better at data computation based on low-level input features. Humans, on the other hand, can serve as analysts, forming high-level hypotheses.
The best way to leverage this is to harness the separate strengths of computers and humans. This enables the two sides to join forces and perform complex sensemaking tasks neither can quickly and easily solve on their own. These include issuing medical diagnoses, performing credit assessments, or determining discrimination cases.
Implementing sensemaking with AI into your business
There are multiple ways in which AI-enhanced sensemaking fits into the world of business as a part of ongoing digital acceleration. It can enhance decision-making, improve performance, and even boost user satisfaction.
Let’s take customer service as an example. As already mentioned, AI has the ability to detect opinions and emotions in written or spoken words. This gives it a chance to absorb client behavioral patterns, improve customer service chatbots, tweak app interfaces, and a large etcetera to best serve the end user. It can also help in marketing and product recommendations, thus potentially boosting profits.
Sensemaking with AI is also extensively used in the medical field, albeit only with supervision from trained professionals. Machine learning-infused services have their use in digital diagnostics for the purpose of early detection and prevention of illnesses, as well as for recommending treatment.
Make it make sense
Sensemaking with artificial intelligence is still (and will likely continue to be) a collaborative process between machines and humans. Gathering information, gaining an understanding of said information, and then using the main takeaways to finish the task requires a human touch. Successful sensemaking requires a clear framework for the machine. With careful planning and targeted implementation, such AI-enhanced sensemaking has its use in science, engineering, and profitable business.