AI’s success story still has a few missing chapters
This article is not written by robots – we promise – but apparently the day when they could be is already upon us.
We have just finished reading about the GPT-2 text generator which, according to this Spectator piece, “not only does text analysis and augmentation, but generates highly sophisticated content”.
“It can effectively write articles,” the piece adds, “albeit slightly peculiar ones.”
The Spectator piece goes on to say that, on account of its ability to generate content “at such speed and scale”, its creators at OpenAI are opting to make only certain aspects public.
“Due to our concerns about malicious applications of the technology, we are not releasing the trained model,” they say.
Malicious applications of artificial intelligence (AI) seem likely to become more and more of a consideration as time passes. Our view on AI at present is more that, while it is unquestionably doing some amazing things, it does not always live up to the claims being made for it by its supporters and in the media.
Read more: This is how the AI crumbles
Faulty towers
One story to catch our eye recently was the news that Japan’s first automated hotel – the Hen-na (‘Strange’) Hotel in Nagaski, Japan, where robots were supposed to carry out all the tasks more usually done by humans – had to ‘fire’ half of its 243 non-human staff.
Among key issues was an in-room robot who was unable to answer guests’ questions and was prone to being activated in the middle of the night by snoring.
On top of that, as this Engadget article explains, check-in robots needed help, concierge bots did not know what they were supposed to be doing, luggage carriers had trouble reaching rooms and entertainment bots broke down.
“Robots aren’t cure-alls,” the piece observes. “Hotels and other businesses have to use them only when they make sense.”
Use AI wisely
That is good advice – not just for businesses but also for those who invest in them.
What investors should bear in mind too is that, while AI may indeed play an increasingly significant part in our lives, companies priced for success on the back of such technology are by no means the ‘no-brainer’ opportunities the steady flow of media stories reverently cataloguing each and every robot success might suggest.
A case study of caution
As a postscript, it seems worth recounting how, researching a separate piece for a big US utility declaring bankruptcy in January, we came across an article talking in curiously detached tones about the company as an investment prospect.
Read more: The story of PG&E and the first climate change bankruptcy
Even more curiously, despite being dated 1 February, the piece never mentioned any of the company’s very public and very recent travails.
How, we thought, could the writer not be aware of the lawsuits, the bankruptcy and so forth?
It took a couple of moments before we twigged the writer was unaware because it had not been programmed to be aware and the piece was one of many being automatically generated these days by brokers looking to drum up business.
Readers – and investors – beware …
- Kevin Murphy is an author on The Value Perspective, a blog about value investing. It is a long-term investing approach which focuses on exploiting swings in stock market sentiment, targeting companies which are valued at less than their true worth and waiting for a correction.