AI Policy Exchange
Let's decide the future of AI, together.
The five AI myths to be aware of
by Anand Tamboli, tech futurist and award-winning author of “Keeping Your AI Under Control: A Pragmatic Guide to Identifying, Evaluating, and Quantifying Risks” (Apress)
Technology is often over-hyped but underestimated. Artificial Intelligence couldn’t have been any different from it.
There are several stories, which don’t give you the full picture. And then there are many anti-stories, which create more confusion. As I often say, the truth can be uncovered only with deep questioning and relentless persuasion for answers. It is simple, but not easy!
So, here are five AI myths that keep circling without any conclusion.
Myth #1: AI will create more jobs than it will eliminate
Whenever there is a question around job losses, most people would play safe and not speak the truth. Sometimes, it is merely due to an inability to see a bigger picture.
AI will eliminate routine jobs, and also it will change many jobs in the coming few years, mainly because AI is currently being used mostly for incremental and efficiency innovations in the businesses.
As Clayton Christensen explained in one of his lectures, efficiency innovation doesn’t create jobs; it instead eliminates jobs and releases capital. That’s where enterprises and many businesses see more value, i.e., fewer overheads and more capital to play with.
AI can help in creating new jobs only if it is democratized significantly, which is not the case as of today. In the future, if the business propensity shifts and more resources are spent on democratization, we will see the change in the job market.
AI can help in creating new jobs only if it is democratized significantly!
However, remember, these jobs will not be like-to-like. That means most people will be required to reset and relearn a lot! Now, those who live hand-to-mouth (on meager salaries or wages); it is going to be challenging. Significant time will be spent on learning new skills whose application may be short-lived. So, the loop of “relearn – apply for jobs – and use those skills for short term” can soon become exhausting for a significant number of people.
Everyone must understand this reality and adjust, merely believing in tales that there would be more jobs (or less) will not be helpful.
Myth #2: AI is smarter than people and will outpace human intelligence
The fact is, AI can be only as smart as you program it. However, in comparison, it may seem more intelligent.
What I mean is, if more people start to lose their wits and don’t learn, keep themselves augmenting mentally and experientially, they eventually will become less and less intelligent in comparison with technology, and the balance will tilt. Mindless use of technology and consumerism are some of the most significant contributors to this effect.
Mindless use of technology and consumerism are some of the most significant contributors to creating intellectual disparity.
One thing we must not forget, there are many dimensions of intelligence. AI knows only a few of them. As long as humans learn to master their other dimensions of intelligence, things would remain in control.
However, if we choose to regress over time, then…godspeed!
Myth #3: AI will solve all humanity’s problems
To solve any problem, you must define the problem effectively and correctly. If humans can’t do that job properly, nothing much will change.
AI may assist in identifying solutions, but humans will have to execute. But most importantly, we will have to define our high priority problems first!
Let’s take an example of the climate change issue. Scientists have been saying it for decades; activists have been pursuing for decades, nothing has changed. Moreover, the environment has given us enough warning signs in the form of several calamities, yet, nothing has changed. Even if people come up with any extraordinary solution using AI, what will change, and why?
We somehow have learned to discard, undermine, or ignore all the solutions we do not like. Much like a bitter pill, there seems to be an inherent resistance not to take it.
The same goes for AI solutions – if we don’t like the answers, we may not accept and execute them. And, even if say we like them for once, we humans will still have to execute flawlessly, go through change management, transformation, the whole nine yards.
And after all, what are humanity’s problems? Customer complaints for which AI chatbots are developed? People identification for which facial recognition is being perfected? Or understanding what people prefer to buy the most for which marketing systems are being fine-tuned? What about real humanity’s problems – inequality, hunger, climate change, discrimination, mental health, etc.?
Coming up with the solution is only 5% of the whole endeavor, we still have to handle the 95%.
Myth #4: Cognitive AI can solve problems in the same way as humans
There is this common fairy tale belief that cognitive AI will help in solving problems in the same way as humans do.
But I hate to break it to you – it won’t!
AI or any other technology, cannot solve problems it wasn’t designed to address. After all, technology is a tool, not the mind. A false belief that it can happen is a pretty intense story.
However, think again! With all this hoopla, how much help AI has done to come up with solutions to the Covid-19 situation? Almost nil!
AI cannot solve problems it wasn’t designed to address.
Myth #5: AI systems are biased
AI amplifies bias mindlessly, which already exists in designers.
By definition, bias means learning from patterns and experiences. Not all biases are bad, wrong, or evil. Neither a weapon is biased nor the AI. However, just like the use of a firearm has a dependency on its user (soldier or terrorist), the use of AI will be subject to who is developing it and using it.
If researchers, developers, sponsors, and early adopters fail to notice or even acknowledge the lacunae of AI systems, it will only perpetuate the existing societal bias.
Although the bigger question remains, “Who decides what is right or wrong?”
A better way to approach the bias issue is to refer to the societal as well as legal fabric and consider it to be the baseline. If you don’t like anything or think it is biased, changing the policy and law will be more effective. Calling out such issues is still necessary and will help to make some noise, but it will not be sufficient to yield fruitful outcomes.
AI amplifies bias mindlessly, which already exists in designers.
The point is
Technology is often over-hyped but underestimated. Remember that people use your ignorance to sell you something! Selling AI revolution is one such attempt. You don’t have to be ignorant and believe in one-sided stories.
Everyone needs to learn to find anti-stories for each story and cross-check. Eventually, you may still believe in the main story, but you will be more confident as a result of cross-checking.
Technology is often over-hyped but underestimated. Remember, people use your ignorance to sell you something.
Views expressed above belong to the author(s).
© 2019-23 AI Policy Exchange
AI Policy Exchange is a non-profit registered under section 8 of Indian Companies Act, 2013.
We are headquartered in New Delhi, India.
Founded in 2019 by Raj Shekhar and Antara Vats