In the battle between AI vs human, one vital thing AI can’t give you is first hand experience. When it comes to advice or recommendations in particular, the benefit of real experience, can make it much more valuable. In my own career, human recommendations proved to be life changing. Maybe that’s because AI has only begun to make an impact but there is no denying the power of a wise recommendation. If I hadn’t been offered a first hand recommendation to study Electronic Engineering, I probably would have followed a very different path. In this article we’ll go through the benefits of human recommendations vs AI based ones and discuss the differences – at least based on the capabilities we have today.
Making the case for humans, first hand experience gained through direct observation has some unique advantages. In the form of a personal recommendation, sharing first hand knowledge can provide genuine insight. There is undeniable wisdom that comes from a direct experience. It can only be expressed from a person going through that event or observing it.
When information is only second hand and hearsay, it’s best to be weary, in this case especially if it comes from humans. We know that when information is degraded through vague recollections and generalizations, it is unlikely to be correct. But when we come across good advice from a human, it’s often because it originates from a real experience. It’s most likely backed up with an assertion – “When I did X’ ” rather than “I heard about X’ engendering greater trust. This is the type of first hand recommendation that give humans the edge.
In the case of computers, advice is offered based on what was ‘heard’ from a human or a from large sample of humans. Computers can tell you that x% of people got this result when they did that thing. It’s usually very effective for straightforward decision making. If you need a yes or no answer with a probability attached, computers will give you superpowers in this respect. You’ll know which route to travel, maybe even what investment to make, anything where likelihood based on a large population results in the sharpest answer. However for those tricky next level decisions like career decisions or what to buy that isn’t just based on popularity or best sellers, your best bet is a human. If you want to know that a recommendation is coming from true, honest experience, it can only come from a person. As long as that person isn’t being paid to make a sponsored recommendation of course. AI algorithms will present you with options based on many factors and they are getting more and more accurate every day. One day all recommendations could come from computers and they may be even more powerful than a single human recommendation. But we aren’t there yet.
With today’s technology, most recommendations are similar for everyone and made up from broad strokes that only apply small amounts of personalization. Some of that is because we are naturally concerned about giving too much personal information up to big companies. It makes sense that there will be a limitation to the quality of the recommendations we get if we aren’t providing the personal information necessary to improve algorithms. It’s difficult to see how this would be resolved until we start sharing a lot more personal data with computers.
So what’s the benefit of first hand experience over AI?
The answer could lie in the way we evaluate a recommendation or piece of advice. To some extent whether a recommendation is valuable or not, comes down to how we choose to accept it.
The whole business of taking advice and recommendations can be an art in itself. What is right for one person, can be totally wrong for another. The subtleties of decision making based on recommendation can be tricky. With so much information to hand, it gets even harder to sift the noise from the useful signal. The power of a first hand experience, usually presented in the form of a story, can cut through the fog. When a human weaves in a real story, a narrative of why they are making a recommendation, it just has a bigger impact. Computers can’t make that up. They can gather up other people’s stories – if they can obtain them but they can’t create them (at least not truthfully). Distinguishing one story from another could be a challenge for computers. So many little details in a story can be a factor that a recipient of the story might struggle themselves to explain why one story stands out among others.
Aside from how we personally accept a recommendation as humans, there are other factors to consider. A human brain has the ability to blend together an enormous number of data points to come up with a conclusion. Not just what is being said or done, but subtle gestures, emotions, anticipated reactions, even taste or smell – so many things that a computer is still learning to do. We haven’t yet invented a computer with that level of complexity. Everything is based on logic and probabilities. But we know humans are just not like that. When humans recommend or give advice (or do any interaction for that matter), they weigh up a lot of things – whether consciously or unconsciously. Their brain is calculating a decision in a far more sophisticated way than any computer available today. That’s not to say we won’t get there eventually – who knows. Somehow, our brains are designed in a remarkable way to digest all the different data sources and within seconds, react accordingly. It’s amazingly efficient at it too unlike the massive amounts of data and training that computers require just to make basic assumptions.
Lastly, there is that mystical factor – call it intuition, call it a sixth sense. It’s often apparent when there is a lack of data available. Humans can fill in the gaps with their intuition, common sense, and gut feel. It’s something we know is not just a result of many years of built up knowledge because babies have it too. We just know some things like they are programmed from birth. We don’t know how to program that into computers because we barely understand it ourselves. Reading about ‘reason’ (you can find my post inspired by the book “The Enigma of Reason’ here ), tells you that it is still largely unknown – why we do what we do is not rational or logical but it works. This intuition is added to the mix when a human makes a recommendation. They can include all the logical reasons for giving you the benefit of their wisdom but the addition of ‘intuition’ makes it unique and unrivaled by computers. ‘I think it’s best to do X’ makes it sound a lot simpler than it really is. Beneath the surface, all kinds of complex calculations with intuition, a good narrative and all senses involved, combines together to produce a human recommendation.
What about covid’s impact on first hand experience?
One thing to ponder is that the arrival of Covid could mean we have a dramatic reduction in the amount of real experiences occurring – at least physical ones. There is an the absence of direct observation which means we can’t be there so we’re reading about it or watching it instead. What impact will that have if any? Maybe it will be a temporary pause that will soon be forgotten. But it could present gaps in the overall human recommendation engine. People will have less to offer each other in the form of first hand advice, if we aren’t able to experience it in the first place. Covid is one more reason to seek and treasure good, human sourced, first hand recommendations.
We built a place for first hand experience
Here at Recknsense, we started recommendation lists for that very purpose. You can look for recommendations anywhere on the internet but they are buried away in feeds or forums. The concept here is to have a list that lasts forever. One place to find recommendations and one place to ask for them. It’s early days. We have one list asking for book recommendations (what book influenced your career?) And only one path to request (the editor!). Eventually it would be great to allow people to ask for their own recommendation lists as well as add to others.