Re: Who'd submit to the benevolent dictatorship of GAI anyway?

From: Adrian Tymes (wingcat@pacbell.net)
Date: Thu Sep 04 2003 - 15:44:49 MDT

  • Next message: Adrian Tymes: "RE: Who'd submit to the benevolent dictatorship of GAI anyway?"

    --- Brett Paatsch <bpaatsch@bigpond.net.au> wrote:
    > In the seed AI scenario, someone will initially made
    > the buys
    > or sells on the AI's behalf as it won't be a legal
    > person and
    > presumably won't be able to trade. This means the AI
    > is
    > likely to be treated as owned by the proxy and the
    > proxy's
    > friendliness rather than the AI's would seem to be
    > the point.

    I think you missed a small but important point: while
    it is true that all of this is *legally* happening
    under the human proxy's name, that is but a legal
    fiction. In truth, the proxy need not be aware of
    the fine details of what actions the AI can directly
    take, for example issuing buy and sell orders through
    an online daytrading account. There are times when
    the proxy's active assistance will be necessary, but
    the AI could minimize those if gaining the proxy's
    cooperation would prove difficult, for instance if
    the proxy's and the AI's goals diverge.

    > The first AI would have to buy the stocks through
    > proxy.

    Or through an account the proxy sets up. Once the
    proxy has signed the right forms, relying solely on
    electronic communication - perhaps spoofable, perhaps
    signed with a key the proxy shares with the AI -
    might well be the expected method of operation.

    > The wealth and power go back not to the AI
    > (and non person under law) but to its human proxy.

    This is true from a legal point of view. But again,
    so long as the money and stocks remain in accounts
    the AI has access to, the AI can manipulate these in
    the proxy's name.

    Of course, the proxy can, at any time during this
    phase, shut down the AI's access. This is one of the
    critical weaknesses of the AI, and avoiding it might
    therefore well be the main goal of any interaction
    the AI has with the proxy.

    > But how would the rest of the world react. AI's
    > human
    > proxy Joe is getting super rich wiping out his
    > competitors
    > and outcompeting everyone with a rate of change that
    > they
    > can't match. This looks to the losers like an arms
    > race or
    > marketing war against Joe who is using the AI as a
    > tool.

    This is true, and likely given human nature.

    > Politically it is very hard to see people standing
    > for it
    > and not attacking Joe or trying to counter Joe and
    > his
    > pervieved "tool" AI.

    It is likely this would happen. It has happened in
    the past to people who got rich using tools they
    manipulated far more effectively (at least, to the
    end of getting rich) than other people. I submit
    Bill Gates as a recent example - and he still has his
    fortune at this time. (Argue back and forth as to
    whether he deserves it, the effectiveness of
    Microsoft towards other goals, et cetera and so
    forth. The fact remains: he's rich, and it's
    unlikely he'll be non-rich any time soon.) It seems
    likely that history would repeat itself, as it has
    before.

    > > A bit more capital would
    > > allow it to become the sole stockholder in at
    > least
    > > some of these cases, streamlining the process.
    >
    > Not it Joe, it proxy. It can't own shares its not a
    > person.

    It/Joe, but if the company only interacts with Joe
    via email and telephone (thus, actually interacting
    with the AI) and comes to think of Joe as an
    eccentric billionaire (which might not be far from
    the truth at this point), what difference does it
    make?

    Or consider what happens if the AI has a very
    lifelike robot replica of Joe constructed, designed
    to fool biometrics, and then disposes of Joe's body.
    (Kill, perhaps. Perhaps just encourage to move far
    away from society such that no one else ever sees him
    again. Same difference, mostly, though the two
    scenarios have slightly different risks.) There is
    then no longer any difference between Joe's goals, as
    perceived by the rest of the world, and the AI's
    goals.

    > Yeah but what about the people that see the AI using
    > Joe as his buyer (and Joe is really getting rich and
    > powerful btw and Joe is not designed to be friendly
    > in all folks terms). My point is Joe and his AI are
    > likely to face a strong political backlash.
    >
    > Aren't they?

    Yes. But money tends to be able to buy political
    influence, or at least gain political influence from
    those hoping to tap into the wealth.



    This archive was generated by hypermail 2.1.5 : Thu Sep 04 2003 - 15:57:45 MDT