It's not a large problem and that's exactly why the high frequency trading industry is tiny. You only think it's big because it gets disproportionate attention in the media.
I guess it depends on how you bracket the problem and how you define "large". According to http://www.marketswiki.com/mwiki/High-frequency_trading,
"It was estimated that as of 2010, high-frequency trading accounted for an estimated 50 to 70 percent of trading."
Yes, that's absolutely natural. That's like saying 50 to 70 of Internet traffic went through an ISP.
I suppose. I mean, I get the difference between quality and quantity, but if we're talking about the gravity of the problem, the knowledge that 50-70% of trades are HFT/automated tells us something about the nature of our securities, commodities, etc. markets. Doesn't it?
But why does that astound you or strike you as a problem? Look at any mature market and you'll find middlemen buffering supply and demand. Imagine:
50-70% of auto transactions are between dealers who don't even drive their cars and only own them for weeks!
50-70% of bread transactions are between distributors and supermarkets who don't even eat bread and want to sell it in days or even hours!
That's a good point. I suppose I assumed that the 50-70% figure referred to front-side securities, futures, or commodity transactions (the equivalent of an end-customer buying a finished product in a retail transaction), rather than amongst various intermediate levels in a complex supply chain.
What in your opinion would be the contribution of a high frequency trader to a society or an individual ? Getting markets to be arbitrage less at sub-mili-second time scales was a pressing problem for no one.
For some of the other industries, if you price the externalities, it is far from clear if they are providing a benefit.