Friday, 1 September 2023

Let's See Them Argue This One!

The former head of the MI6 who was in charge during the UK's invasion of Iraq has claimed scientists' warnings about the dangers of artificial intelligence can't necessarily be trusted because of incorrect claims they made about Covid-19
Sir Richard Dearlove said that as 'brilliant' scientists had gone 'off piste' during the pandemic, he was sceptical of experts telling him AI will destroy humanity.

Oh boy, get the popcorn!

This should just be enough...
 

'They actually used their knowledge and speciality to influence government policy.
'So at the moment, I am in a mood to be quite sceptical when some brilliant scientist says to me "I know all about this and I am going to tell you authoritatively what is going to happen".'
He said he reacts with a degree of cynicism when scientists share extreme information in regards to AI.

I react with 'a degree of cynicism' these days to everything! 

7 comments:

  1. You can reason from first principles that the go-ahead for AI will turn a 'scientific' blind eye to all risks and hazards it involves. At the time of the Manhattan Project, the possibility of creating fusion conditions in the Earth's atmosphere was a known 'small' risk which wasn't allowed to impede the test. One way or another, the nature of Greed will be pivotal in the premature destruction of the planet.

    ReplyDelete
  2. I (being an avid sci-fi reader) almost believed the hype but then I ... tried using multiple 'smart' search engines to identify a music video I remembered (to a human it was quite memorable and unique, but not apparently to computers).

    These are software systems constantly 'improved', refined and refined again for decades remember. The very pinnacle of software excellence and integration.

    After three weeks of every possible way of describing it, and every permutation of language to do so, I gave up. Then I mentioned my difficulty to someone standing in the queue at the coffee shop and ... the guy standing behind me identified it from the vaguest description possible.

    I've worked in industry and healthcare, both use multi-billion pound developed 'expert' systems that work brilliantly in day-to-day standard situations, but give them one small variance and the system collapses completely (with some revealing and always humorous, at least if caught before the product hits the patient, results).

    We may (have) AI beat a chess master, but AI coping with the ultimate in chaotic systems (otherwise known as the real world and worse ... people) ... not a chance (at least not for 'cold fusion' six months - meaning possibly never).

    Throw in the fact that some prepubescent nerd 'will' end up hacking it and ... I doubt we'll have to worry about Skynet for a while (although HAL is a distinct possibility - following rules some corrupt or inept programmer included without thinking, and thus ... "Open the garage door HAL I'm late for work").



    ReplyDelete
  3. As we saw in the Covid event, when so-called scientists start to make projections, it makes astrology look almost respectable.
    The same applies with economic forecasts and man-made climate-change hysteria - I'd rather believe Mystic Meg.

    ReplyDelete
  4. Politicians need no help from anyone to bugger anything up.

    ReplyDelete
  5. AI doesn't exist yet and there is an argument that it'll be a long time coming.

    What we have atm is Machine Learning and brilliant coding. Compounded by the fact computers are so much faster. So very smart people make them look intelligent. That only works so far and Biden is a perfect example that you can make things look intelligent forever.

    As far as science goes most of these people are not scientists in the area they talk about. A doctor with a doctorate in medieval flower Arranging talks about climate change end everyone things because they are a Dr they know what they are talking about.

    Nobody with any sense is going to believe scientists for quite some time.

    ReplyDelete
  6. "...the go-ahead for AI will turn a 'scientific' blind eye to all risks and hazards it involves."

    Probably, yes. We never learn, even if the machines eventually do...

    "Then I mentioned my difficulty to someone standing in the queue at the coffee shop and ... the guy standing behind me identified it from the vaguest description possible."

    Thank god for the human response! We aren't obsolete yet...

    "As we saw in the Covid event, when so-called scientists start to make projections, it makes astrology look almost respectable."

    We'd have been better off right now if she'd been in charge!

    "Politicians need no help from anyone to bugger anything up."

    Good point!

    "AI doesn't exist yet and there is an argument that it'll be a long time coming.

    What we have atm is Machine Learning and brilliant coding. "


    Mostly brilliant coding...

    ReplyDelete
  7. "...the go-ahead for AI will turn a 'scientific' blind eye to all risks and hazards it involves."

    Probably, yes. We never learn, even if the machines eventually do...

    "Then I mentioned my difficulty to someone standing in the queue at the coffee shop and ... the guy standing behind me identified it from the vaguest description possible."

    Thank god for the human response! We aren't obsolete yet...

    "As we saw in the Covid event, when so-called scientists start to make projections, it makes astrology look almost respectable."

    We'd have been better off right now if she'd been in charge!

    "Politicians need no help from anyone to bugger anything up."

    Good point!

    "AI doesn't exist yet and there is an argument that it'll be a long time coming.

    What we have atm is Machine Learning and brilliant coding. "


    Mostly brilliant coding...

    ReplyDelete