All over the world in boardrooms great and small can be heard talk of artificial intelligence and machine learning. AI/ML have become the short form way of writing about it because the words are used so often, it actually reduces productivity to write it out. They are no longer just words, they are buzzwords, power words. They will be mentioned in every official communication from every company. It doesn’t matter whether the subject is manufacturing or the driest pages of a quarterly report. CEOs are extremely concerned about whether their machines are learning.
That’s fine as far as it goes. Machine learning is powering a lot of what the company will be doing for the foreseeable future. However, one suspects that some executives used the terms without having a strong grasp of what they really mean. No one wants to seem like they are not keeping up with the latest trends. But not every machine needs some form of artificial intelligence to be useful. No one wants an artificially intelligent hammer that makes decisions about whether or not it will pound nails. Consider these machines that are intelligent enough just the way they are:
There are many surveillance robot applications that don’t require artificial intelligence. What they require is the best piece of fine-tuned equipment money can buy. Your surveillance robot needs to be able to do the following:
- Provide situational awareness for tactical operations, detecting hazardous materials, and relaying critical intelligence.
- Offer a unique opportunity to keep first responders out of harm’s way by letting them check readings and collect data from a distance.
- Enter areas previously considered inaccessible for a wide range of industrial and infrastructure inspections.
You need tech that will carry out your orders and convert your input into real-time action. With machine learning, you can give a partial or imprecise set of commands that the machine then figures out how to perform by understanding what you intend. But when it comes to inspection and surveillance robots, you know what you want and say exactly what you intend.
Science is a precision task that requires precision instruments. Those instruments should only be wielded by highly trained humans who know exactly what they are doing. It is not a field where the human can afford to second-guess the equipment. And the last thing they need is for the equipment to second-guess the human. The scope of artificial intelligence and machine learning is quite broad. It can be found in all manner of industry, factories, retail, education, and much more. But we should be wary of bringing it into the lab.
This is not to say that scientists should make no use of artificial intelligence and machine learning. With the use of these tools, scientists might be able to model certain behaviors of viruses and pathogens. If they can gain some insight into how certain cancers form and propagate, it will be easier to come up with better treatments, perhaps even a cure. Physicists use these tools to peer into the very beginnings of the universe and are able to better predict the movements of asteroids and other dangers to the planet. But at the end of the day, the tools they use to make measurements and that enable direct experimentation to need to be free of intelligence so they can be better used by the people who have it in abundance.
It makes sense that artificial intelligence plays a part in a nation’s defenses. But the UK government wisely plans to ban fully autonomous weapons. We can only hope other countries have the same commitment. No good can come of a weapon that can think. We don’t need weapons to think for our police officers, soldiers, and peacekeepers. We need people who can think better, clearer. While intelligence can be deployed in guidance systems and the like, it should never be deployed in a way that allows the human to recuse herself of any part of the decision to fire and what exactly that person is targeting.
We need machine learning to advance and be deployed wisely. But we will always need machines that simply yield to the natural intelligence of the humans who use them. Surveillance robots, lab equipment, and offensive weapons are best when they are passive tools that are used with precision and care.