The Future of Truth Telling

Telling the truth.

How easy is it going to be to detect a lie?

In the future, we will have artificially intelligent machines. This means some machines will be more intelligent than some humans. It means machines may apply logic better than most humans.

It does not mean the machines will have “feelings.” It does not mean machines have creativity. It does not mean machines will have imagination. It does not mean machines will have empathy. It does not mean machines have understanding. It does not mean machines have the ability to interpret nuances and subtle human interactions. It does not mean machines become human.

I saw part of a TV report which said that in the future we will have artificial intelligence truth checking devices, and in these speech will be instantly translated into text and checked against a data base, to ensure truth is being said.

On the face of it this sounds great but I have some reservations. For example; who enters the data the AI machine checks the words against? For example many facts in history are disputed. The Romans, while trying to subdue the druid lead Celtic tribes of Britain, claimed the druids sacrificed human babies. There is no evidence, as far as I know, other than the written claims by some Romans. The Celtic people at that time had little or no written records. This is not a verifiable truth, it is an unopposed claim but nothing more. If the data used to check truth, has included the Roman claims, a declaration that the druids did not sacrifice children, will be fagged up as a lie.

How will this machine cope with time zones? for example I say it is 12 noon which is accurate for where I am but if the machine is on New York time, it will say this is a lie.

Will people have a truth testing app on their smart phones and check every word said to them for truth?

Will rival companies have different data bases? Will Microsoft, Google, and even Amazon sell truth testing apps that have their own company data base? If so this is going to make dating “interesting” and I would place a large bet on an increase in the number of divorces around the world. How about going into a shop and asking, is this cream cake fresh? Since the term fresh is relative, then how is the gadget to decide what is fresh, is it 2 hours or 2 days old and so what is a truthful answer? Consider the classic question, are you honest? Most of us are honest most of the time, sometimes we tell lies believing them to be truth at other times we tell lies to spare someone else pain or stress. How is the truth gadget to determine the accuracy of the answer given? The present technology is aimed at the spoken and written word, if body language and facial changes, get added then the machine will take over in the law courts.

How about false claims of crime, one person claims they were attacked by another but there is no physical evidence, so how does the truth machine check which version of the event is honest?

How will we ensure that what is put into the machines memory is actual evidence based truth and not someone's belief or opinion? Many people are utterly convinced of the accuracy and truth of a statement, but are still wrong.

Used in the world of politics or media reporting of politics, it will cause chaos, some of the more exuberant political figures use wild exaggeration knowing the audience will understand this is what is being done, but will a machine simply declare the words false. Will all speeches have to be subjected to scrutiny from legal experts, diversity experts, sexual equality experts and from a truth machine. It is going to be a very boring speech, if it is ever made.

The media often chooses to preface a statement with words such as, “allegedly,” or “in some opinions,” and by doing this they avoid legal challenges. This way it does not matter whether a truth seeking machine says it is a lie or not. People will still have to make up their own mind as to the validity of the opinion being voiced.

It may make the media more circumspect in pronouncement of opinions as if they were factual, but I will not hold my breath.

Statistics can be used so support a claim of accuracy but any one who has worked with numbers knows you can use the same set of statistics to justify a different claim, simply by asking a different question or arranging the numbers differently. How is the truth machine going to cope? It is widely known that the selection of which set of statistics to use, alters the conclusions.

Combining robotics with artificial intelligence is going to drastically change the working lives of humans, vast numbers of jobs presently done by humans will be done better, faster, cheaper, and continuously 24-7 by machines that learn. Manufacturing and administrative tasks will end up being done by machines but that does not make the machines human. Artificial intelligence will be developed that can learn factual information faster then humans but it is questionable as to the extent to which such a machine could learn to interpret the nuances of human interaction with large numbers of other humans, will they ever learn how to tell when someone is making an assumption, using their imagination, telling personal lies, or trying to make a joke?