healthHackers

View Original

Episode 36: Blake Richards


Why we forget things, and the ‘realistic’ threat from robots

See this content in the original post

healthHackers® Episode 36 with Dr Blake Richards


No time to watch the video? Below is the Soundcloud audio version or get the podcast on your iPhone here.

See this content in the original post

In this video, you’ll discover:

  • Why we forget things

  • How our brains ‘prune’ our memories

  • Key tips for when you’re trying to remember new things

  • Why you have to take every story anyone ever tells you with a grain of salt

  • Why some drugs can change how empathetic you are

  • How people with perfect memories aren’t more successful than the rest of us

  • What false memory syndrome is

  • Whether computers really will outsmart humans on every level one day

  • The dangers of artificial intelligence (A.I.)

  • What writers and movie makers fail to appreciate about A.I.

  • The most mind-blowing thing Blake has realised about the brain during his neuroscience career


Dr Blake Richards specialises in neuroscience and A.I.

Machine learning, memory, and mind-blowing brain power

Robots won’t rise against us and take over the world, but ones that can kill without hesitation, are “an all too realistic scenario,” according to neuroscientist, Dr Blake Richards.

What sets Blake apart from other academics in neuroscience (aside from his prestigious awards) is his expertise in artificial intelligence (A.I.), which is the imitation of human intelligence in computers and machines.

Most of us are used to media hype surrounding imagined dangers of a robot revolution, but Blake told me in healthHackers® episode 36, that “the danger is not a Terminator scenario where the robots gain consciousness and then decide to kill us all.”

According to Blake, it’s hard enough to program an artificial intelligence just to recognise the difference between cats and dogs

“What many of the writers and movie makers fail to appreciate when they write these scenarios is that, at the end of the day, we program these agents. We decide what their goals are, and as long as we don't explicitly wire into them [that] destroying humanity is one of their goals, that is a very unlikely scenario,” he said.

According to Blake, it’s hard enough to program an artificial intelligence just to recognise the difference between cats and dogs.

“The idea that you're then going to get something that is so capable that it can outsmart us and figure out how to get its own energy supply, and its own weapons, and outwit all of our militaries, and avoid our bombs -  it's a bonkers scenario.”

A more likely threat is the potential military use of artificial intelligence agents to kill on the battlefield without having to think about what they’re doing.

“To some extent, even without A.I. being in the loop, we already see that with the use of drones for example,” he told me.

The brain likes to ‘prune’

At the time of recording our interview, Blake is based at the University of Toronto, Canada, where he specialises in understanding intelligence and memory.

Recently, he and a colleague carried out a review of the literature on ‘forgetting’ and found significant evidence that “the brain actively forgets things.”

One of the newspapers in the UK reported on the review with the headline: ‘A poor memory could be a sign of intelligence, says new University of Toronto Study.’ 

But according to Blake, that wasn’t the best description of his findings. 

“No, forgetful people are not necessarily more intelligent,” he told me, but what he did find was that…

“When we forget things, it's not simply that the mechanism by which we remember things fails, but rather, we actually have some kind of pruning of our memories.”

Blake and his colleague discovered some general reasons that could explain why the brain would invest energy into deliberately forgetting some things.

One of those reasons, is in order to avoid too much “interference,” which refers to those moments when you recall the wrong thing. 

For example, you can’t quite remember someone’s name but you feel like it’s on the tip of your tongue - that’s interference.

“One of the easiest ways to avoid it is simply to erase previous information if it's no longer necessary,” he said.

The brain invests energy into forgetting some things, according to Blake

Blake’s conclusion is that a little bit of forgetting is not necessarily a bad thing.

Watch episode 36 to hear his top tips for when you’re trying to learn or remember something new.

And by the way - there are people out there with perfect or near-perfect memories, although only a handful, Blake told me.

“What's interesting is that these people are not more successful than the rest of us. In fact, they described their condition as ‘frustrating’ and they tend to have symptoms that sometimes resemble OCD. They really don't seem to have any substantial benefit as a result of this perfect memory.”

What has surprised Blake the most during his neuroscience career?

“The more you actually study the brain, the more you realise that the simple things we take for granted are in fact mind-blowing.

“Our ability to literally just walk and talk at the same time is phenomenal. Honestly, it seems easy to us because we are built to do that and we learn how to do it very easily, but at the end of the day these sorts of simple tasks are incredibly difficult,” Blake told me.

“Our ability to literally just walk and talk at the same time is phenomenal,” Blake told me

Blake explained that sometimes we can find certain tasks difficult because we simply didn’t evolve to do them naturally.

“Everyone talks about how hard math is, but math is actually easy, math is a lot simpler than walking. Only our brains weren’t made to do math, our brains were made to walk.”

I asked Blake if a computer could ever be as intelligent as a human brain.

“At this point in time, there's no reason that we have ever been presented with, to suspect that there are any things that our brains do that a computer couldn't do, if provided with the appropriate mechanical recipe for transforming the inputs to the outputs in the way that we do,” he told me. 

I argued that surely a computer could never be empathetic.

He said: “Well, why not?”

In Blake’s view, empathy is a mechanical process whereby you might recognise someone is upset, and you take that stimuli and act on it.

“Empathy is something physical, something mechanical embedded in our brains, and if that's the case, then theoretically if you figure out what exactly that operation is you could replicate it on a computer,” he said.

I’m suddenly rethinking his assurances that computers could never outsmart us.

Follow Blake on Twitter