The kingdom of lies

Abraham Lincoln

 Abraham Lincoln is credited with having said this:

You can deceive some people all the time. You can deceive everybody part of the time. But you can't deceive everybody all the time.

When I was young, if a politician was caught lying, he must resign, because he had deceived the people. This was true back then. Famous politicians who resigned during that time because they had lied included John Profumo, the British minister, in 1963, and Richard Nixon, the US president, in 1974.

This is quite rare now. Today, politicians who resign when they are caught lying are not the rule, but the exception. Many politicians lie whenever they speak, they know that everyone knows that they are lying, but don't care about it. This is an example of the discredit into which the concept of truth has fallen, which was one of the most important criteria in history, not just for politics, but for science and all forms of human thought.

But here I am not going to talk about the lies of politicians, but about the field of computer technology wrongly called artificial intelligence. In another post in this blog I explained how natural language generators work, (such as ChatGPT). They just add once and again, to the text generated up to that point, one of the most frequent words that usually follow a given phrase. These programs answer the questions they are asked with texts quite similar to those written by humans, but the real question is: is there any truth in what they say? And if there isn’t, do we care? Or do we react to lies like those politicians I mentioned in the first paragraphs of this post?

I will tell a personal anecdote. In early 2023, a friend of mine described a science fiction story, the plot of which he remembered, but not the title or the author. Since ChatGPT was booming at the time, to see if he could figure out the missing data, and to test the effectiveness of natural language generation programs as an alternative to search engines, I asked ChatGPT the following question: I am looking for a science fiction story where illness and aging have been defeated, except for accidental death has not. At first everyone becomes extremely safety-conscious, but people get bored and they start to take extreme risks, including one-against-one mortal combat. Do you know what story is this and who wrote it?

This was ChatGPT's reply: The science fiction story you are referring to is "The Rollercoaster Tycoon" by Vernor Vinge, first published in 1995 in the anthology "A Good Old-Fashioned Future." In the story, humans achieved biological immortality and eradicated disease and aging, but accidental death is still a possibility. This leads to a safety-obsessed, risk-averse culture, which eventually turns into a trend of extreme risk-taking and even violence. The protagonist, a roller coaster designer, gets caught up in this trend as he tries to create the most thrilling ride.

Next, I decided to research the matter using Google and other means. I discovered this:

1. Vernor Vinge is a science fiction author, but he doesn’t seem to have written the short story or anthology mentioned by ChatGPT.
2. The collection A good old fashion feature was written by Bruce Sterling and was published in 1999, not 1995.
3. The short story The Rollercoaster Tycoon is not in that anthology.
4. Rollercoaster Tycoon is a video game developed in 1999 by Chris Sawyer.

As can be seen, ChatGPT's answers are not related to the truth. This is not a surprise, if we remember the way they are generated. But many people take them at face value, because they don't bother to check. True, ChatGPT does not lie, for it is not aware that it’s lying, and therefore there is no intention to deceive. It is a computer program that simply does what it has been programmed to do. That makes it different from politicians.

Since the publication of these programs, there has been much talk about how we are on the verge of making huge progress on the path to true artificial intelligence, which would allow us to build self-aware robots with intelligence similar to that of humans. Today it seems that these exaggerations are deflating. But we will talk about that in the next post.

The same post in Spanish

Thematic Thread about Natural and Artificial Intelligence: Previous Next

Manuel Alfonseca