The idea in a nutshell: Most people are focused on existential threats from AI at the moment. That includes me. I feel like I have banged on about them. However, I believe there is a rising body of evidence to suggest that the threat from bioengineering is an even more significant threat. In some ways, it’s worse. Bioengineering seems to get less news coverage, is accessible to more people and is happening much more quickly.
The scary kind of AI is not happening yet
It is true to say that AI has already happened but if you just read that headline, you’d miss the point. AI seems likely to take over the world in 2017 for two reasons:
- There is no clear path to it: The scary kind of AI, the AI which people worry will take over the planet (general AI) rather than take a bit of their job, is a long way away. Scientists don’t agree on how far away it is. Some say 25 years, some say 200. There is no clear, visible path to it either. In his book, Nick Bostrom suggested that emulating the human brain in computers would be the most likely successful avenue, and people are doing that, but no one has said with confidence this is how we will build a general AI.
- The barriers to entry are very high: If you want to be world class at developing AI (and you’d have to be to build a general one), you need corporate resources like one of the big boys. You’ve got to be an Apple or a Google to pay the $400,000 a year for each recently graduated AI geek and to set them playing with your vast data stores. Vast data stores tend to be facilities held by big companies, too.
But the opposite is true of bio-engineering
Unfortunately, there is a path to success in bio-engineering and barriers to entry/production are low. Bio-engineering (which, as I understand the term means changing elements of genes to enhance something living in a direction you want) does have a clear path to progress.
New bio-engineering stories with incredible consequences are released almost every day :
- At the centre of it all ( from 2012 ) appears to be CRISPR: This TED talk explains how scientists can use a technology CRISPR to change a gene and then insert it into a body’s DNA using a cleverly adapted virus. The result could be that in the future children are both with stronger limbs and without genes which cause baldness, for example.
- It is possible to make a genetic change yourself for around $1000. You can invent a new stream of a ‘pox’ for $1000 by inserting a gene into a virus. The technology is widely known which makes it a relatively easy tool to be used by foreign governments. This article from 2002 raises the threat posed by Iran, Iraq and North Korea, all of which are thought to have started engineering their own smallpox stores. Smallpox is considered ‘by scientists’ to be ‘the pathogen most dangerous to the human species.’
- So, some people have started experimenting themselves: In this example, a man seeking to extend his life is investing in some personal gene therapy. This particular application of the technology is so likely to come to fruition that the article describes him as a visitor from the near future. This time using an electric current to open his genes for (a more temporary) adaptation, the subject inserts genetic structures (bought online for around $10,000) into his legs which he believes will make him stronger.
- And that’s not the only example of this stuff happening without proper regulation: It was also recently reported that a new discovery (in vitro gametogenesis) means any type of cell can now be reprogrammed as a sperm or egg. The article had a real-life attention-grabbing example of a potential real-world ethical dilemma:
Imagine you are Brad Pitt. After you stay one night in the Ritz, someone sneaks in and collects some skin cells from your pillow. Next thing you know, a novel fertility technology has used your movie star cells to create sperm and make a baby. And now someone’s suing you for millions in child support.
So this chart is a bit misleading
In this chart, from the WEF, bio-engineering is seen as having almost equally catastrophic consequences as general AI. What the chart doesn’t show is how much further progressed this technology is compared to the AI alternative.
Missing the timeframes of the relative threats does appear to suggest AI is a bigger threat. We may all be wiped out by a basement generated Brad Pitt Pox before general AI becomes a threat, however.Summing up
The problem, as many have said, is that the benefits of bioengineering are tremendous. They may make the lives of even adults better, helping us live longer, fitter lives and providing targeted drugs which I have written about in another article.
There are some great potential health benefits: They’ve made a form of the virus Salmonella which eats brain tumours, for example, this article.
Usually, you have to take the rough with the smooth, the risk with the good bits of the new technology. However, in this case, being able to cook up a new strain of a deadly virus for $1000 in a basement, whether you are a person with bad intent or a government with, erm, bad intent, is genuinely of more concern to me than a theoretical threat from AI.
In some ways, bio tech is much better regulated than AI at the moment when it comes to companise developing things for human consumption or drugs. Even for animals. As one of the articles above suggests, you need to fill in a form and get it signed off by the authorities before you develop new gene therapies. Unfortunately, that regulation can be ignored when you can do to for $1000, as we’ve seen with the guy operating on himself to extend his life. And we all share the consequences of a mishap.
A very smart bloke, Andrew Fisher put me on to The book The Windup Girl which described the downstream effects of a world in which Bio-engineering got out of the box. I didn’t realize it was all so close.