Re: Can an Algorithm Write a Better News Story Than a Human Reporter?

Re: Wired Magazine 4-24-12 

A company called Narrative Science has developed software that can compete with human authors in writing news stories (see YouTube Video). Forbes is among their many clients. The company believes that its system will evolve in its ability to write stories that will go “higher up the journalism food chain—from commodity news to explanatory journalism and, ultimately, detailed long-form articles”. Narrative Science’s leader says:  “In 20 years, there will be no area in which Narrative Science doesn’t write stories.”

A couple of weeks ago, Education Week (4-12-12) put out a similar story. In this case the computer was competing with college professors at scoring essays (see our post). On 4-26-12 Extreme Tech carried a story (Will an IBM computer be your next mayor?) about the use of IBM’s Watson to manage the operations of the city of Rio de Janeiro.


IBM’S WATSON – Natural Language
AI that won TV’s Jeopardy championship

Smart machines (robots and AI systems) have already taken over tens of millions of manual labor and routine mental labor jobs and they are rapidly developing to the point of being capable of replacing more and more of the ‘thought-work’ of humans. What does this mean for the economies of the future?

Among the the comments following the Extreme Tech articleNattyspats stimulated the conversation with:

The real issue  is what this means for labor and employment policies?  When machines manage autonomously with human oversight, what level of unemployment is acceptable?

In response to the above, the editor of Extreme Tech replied:

Did you miss the whole bit at the end, about such a system replacing skilled workers, and there being no turning back?

Nattyspats replied to the editor with:

That speaks only to the elimination of positions, not the outcome of displaced workers.  Today in the U.S. the nominal unemployment rate is 9%.  So, in an autonomous society is 45% unemployment acceptable?  Plainly it is no advantage to IBM – or anyone else – if there are no employed workers which contribute to 401k plans which buy stock . . . which includes IBM and Cisco.

I replied to this thread and article and built further on our KeyMeme, The Challenge of Change:

Clearly the machines will extend into replacing humans everywhere it is cost-advantageous and socially acceptable to do so.

It might not be advantageous for IBM – or anyone else – to allow the displacement to tip the economy they depend on. However, it will certainly appear that it’s in their near-term best interests to exploit the opportunity. There is a lot of money to be made replacing humans with machines. There is an ever growing list of once exclusively human jobs that are going to become machine-replaceable. Reducing the human costs of business means greater profit. It’s great for the system providers and it’s great for the companies saving money. It’s not problematic for the companies involved until the human displacement effects redistribute the economy beyond the capacity of ‘market corrections’ to re-orient and re-balance.

I don’t trust the ethic or the intelligence of corporations when it comes to this one.

Changing this trajectory towards destabilizing displacement might be impossible. It might take the crash of wealth itself before the necessity of balance redefines corporate ethics enough for a new kind of economy to form. Given our political learning disabilities it’s hard to see how human displacement concerns could be the basis of any kind of helpful ‘regulation’.

For me, the most important conversation begins when we accept  that the displacement implications, whatever they might be, are sufficiently important to ask:

What is the mission of education?

In light of what is coming, how can our education system best steward the learning of children (and the future of the economy) so as to maximize their employability as adults? (among other reasons)

With intelligent machines getting ever better and more cost-optimal at performing what was once exclusively human jobs, what should we be teaching children in school today in order to be ready to compete with machines in the future economy?

I posted the a version of my question on the comment stream of Wired’s article about ‘reporters‘ and got this from JeramieH3:

> teaching children in school to be ready to compete with machines

Unless we’re teaching our children to be the next John Henry, we should be teaching them to complement the machines, not compete with them.

To which I replied:

Excellent point. I agree. I meant by ‘competing with machines’ that the machines will continue to advance in being able to perform what was once humans jobs. Where we meet (I think) is that humans must learn to be good at things that machines can’t do better (more cost optimally when viewed in terms of ROI by corporations). This is but one of many threads on my site that lead to the necessity of rethinking the ‘mission’ of education. Training kids to be inevitably replaced by smart machines is a disservice to them and to the future of our economy. 

Stay Tuned

, ,

No comments yet.

Leave a Reply