Artificial Intelligence Vs Human Decision Making: Do We Have A Winner Yet?

Garima Aneja
Mar 02, 2019   •  8 views

Artificial Intelligence simply means that machines can perform tasks in ways that are “humanly intelligent” through simulation and pattern recognition. A few years ago, artificial intelligence and machine learning was only an idea seen and experienced in sci-fi action movies. But today, AI powered systems are actually solving problems and improving productivity. Byron Reese in his latest book “The Fourth Age”, calls ‘artificial intelligence and machine learning’ the fourth major change in human civilization after ‘discovery of fire’, ‘development of agriculture’ and ‘discovery of writing and wheel’. AI powered systems are yet to achieve their full potential and we are sitting on the cusp of many more breakthroughs this technology can achieve. While artificial intelligence has contributed to some notable breakthroughs in recent times, there is still considerable debate on whether artificial intelligence will replace human intelligence or in other words, are AI powered systems better at taking decisions previously taken by humans.

Is AI better at taking decisions than humans?
When it comes to processing huge volumes of data, human capability is limited and it is in this sphere that computers outperform humans in terms of speed and efficiency of not only data processing and analysis but also identifying patterns in data. Machines can identify patterns when supplied with huge datasets and can develop algorithms to identify similar future patterns which assist decision making in the future, and machines are able to do it quicker and more accurately than humans.

It has been observed that AI powered systems make more accurate decisions especially in areas like choosing which stocks to invest in, since these decisions involve spotting patterns from a huge amount of database and evidence. AI powered systems are highly useful in areas

where a lot of prediction is required such as measuring an asset’s risk premium, its liquidity, volatility and price trends over the past few months or weeks.

Take the medical profession for example. Each new day, there is a constant influx of new information about signs of disease, new treatments and new medical technology. It is humanly impossible for a doctor to absorb all this information and to base their future decisions taking all this information into account. AI powered systems and new medical tools absorb all this information and provide a decision that takes into account all the possible comprehensive data and these machines do that with less error margin and at a fraction of the cost in speed time.

Take the case of a Human Resource Department in a large organisation. Organisations can leverage artificial intelligence in performing repetitive tasks like sorting resumes, interview scheduling etc. at a fraction of human labour cost within record time. Such a system will also ensure efficiency as computers will be able to provide standardised feedback to each application.

Coming to strategic decision making in organisations, AI based systems will improve the quality of decisions made and in turn, corporate governance overall as decisions can now be taken by incorporating all the environmental data sources and facts as opposed to humans taking decisions with limited sets of data. These decisions will also be less prone to emotional or human bias and as a result, will be based on objective facts and the organisational context.

Banking and financial service industry is another case in point where AI powered systems are taking quicker and smarter decisions. Banking areas like data entry, reconciliation of banking statements, accounting software, chatbots etc. are performing the repetitive cognitive tasks in record time. With banks and financial institutions facing increasing amounts of unprocessed data, it is not only costly but also time consuming for humans to process it.

Even in the financial advisory industry, artificial intelligence is being increasingly used to read deposit data for thousands of customers and take important decisions like flagging high payment deposits and sending notifications and sending reminders for specific transfers.

The speed and agility by which artificial intelligence operates has made it very useful for military and national security purposes as well. For example, the USA government has deployed AI through its Project Maven which goes through huge volumes of video footage and data and sends signals to humans regarding any abnormal activity noticed.

Thus, in conclusion, AI powered systems take better, quicker, unbiased and more accurate decisions than humans, especially in areas where there are a large number of data points and where the tasks involved are frequent, repetitive and structured.

The other side of the coin
There is no denying that artificial intelligence and decisions taken by machines are quicker and cost effective, but that does not mean that AI can replace human intelligence or AI is better than human decision making. AI works effectively when decisions are based on analysing huge volumes of data, rules or patterns. But not all decisions involve these elements. There are so many decisions in real life which use creativity, intuition and a human’s tacit knowledge and AI can never replace human intelligence in these cases. AI powered systems can at best support human reasoning and judgement but cannot replace it.

Continuing with the example of medical profession, artificial intelligence and algorithms, however quick they are, lack the fundamental understanding and reasoning of medical concepts and the decisions they take are devoid of understanding of situational circumstances and human emotions.

Take customer service for example. AI powered systems can provide standardised responses to customer complaints but in today’s

competitive environment, people prefer to talk to real people for their problems. Human decisions in this case are more reliable as they value empathy and consider the customer’s emotional needs rather than just considering patterns from data like AI. There is always an argument that decisions taken through AI are free from any human error or bias. But this argument is inherently flawed as decisions taken by AI powered systems are based on data fed into the machines by humans. If the data is erroneous and biased, decisions taken by even the most complex software are also biased. Decisions taken by AI powered systems are as good as data that humans feed them and hence cannot be blindly relied upon. Another problem with AI powered systems is that they are not “intelligent” in the same sense as humans. You supply the system with a huge volume of data and it analyses it and stores the patterns from it for future decision making and application. Thus, artificial intelligence is dependent on data and the more data you feed into the system, the more useful it becomes. But “intelligence” is much more than mere data analysis. AI powered systems are not able to “understand” the data in the same way a human brain can and hence cannot take decisions when the context and scenario is tweaked a bit.

Continuing with the HR Department example, AI powered systems can take efficient decisions while sorting resumes or scheduling interviews but it is not possible for these systems to take decisions in areas where a human touch is required, for example taking an actual interview and designing and implementing training programmes.

AI powered systems are great at performing standard and repetitive tasks based on past data and patterns but cannot replace or become better than human decision making which is crucial for more complex and analytical tasks. AI lacks what human intelligence possesses: common sense, intuition and learnings based on a vast reservoir of experiences. A hundred percent reliance on decisions taken by AI (in the hope that it is superior to human intelligence) will actually run the danger of not accounting for all the hidden costs, risks and rewards which are considered by a human brain when a decision is taken.

Who is the ultimate winner?
Both artificial intelligence and human decision making has its own pros and cons and both artificial intelligence and human intelligence have their own suitability in different circumstances. By saying that AI powered systems perform better or worse than human brain and by rendering them as mutually exclusive, we are narrowing the usefulness of the debate.

Take the simple example of streaming service Netflix. Using powerful AI algorithms, people’s past streaming data is analysed i.e. which show they watched, which genre of shows interests them, how long they watched each show, etc. and based on all this data analysis, Netflix provides a consumer with a list of shows best suited to the consumer’s interest. Now a consumer uses this AI generated information and combines it with his or her own human judgement, the mood he or she has, the context, the time the consumer has etc. and finally makes the decision about which show to watch.

Decisions taken by machines have their own advantages and are suited to different circumstances and the same is the case with decisions taken by a human brain. When it comes to analysing huge volumes of data and predicting patterns for future application, AI powered systems perform better than humans as these systems are faster, cost effective and more efficient. But not all decisions involve making accurate predictions. And this is where the magical human brain comes into play as it is able to use human judgement and weigh outcomes before a decision is taken. Thus, as is argued in all the above examples, be it in the medical industry or banking industry, AI powered systems are significantly better at taking decisions under certain conditions i.e. when they have a lot of past data to perform simulation and pattern analyses on, but at the same time human decision making is better than AI at predicting “unknown unknowns”. Thus, AI powered systems and humans work best and have an edge in different areas.

Looking at the future of this AI vs Human brain debate, the best solution
is to combine the powerful pattern recognition and simulation strengths of a machine with the insight, judgement and tacit knowledge of a trained human brain. There is definitely a need for a partnership between AI powered systems and human intelligence, wherein each does what it does best to increase the efficiency, accuracy and usefulness of decision making in the future. AI powered systems will enhance human intelligence and be assistive to it but cannot replace it. AI powered systems can perform better than the human brain in specific tasks but the world still needs human judgement, human experience and intuition while taking a number of decisions. Decision making in the future will be better if the horsepower of an AI powered system is combined with the intelligence of a human brain.