As I have been thinking about Artificial Intelligence (AI)
and the interaction of technology and magic, it seems time to review two more
magical objects from the world of Harry Potter. (For an earlier example, here’s
a post about the Marauder’s Map.)
The Quick-Quotes Quill makes its first appearance in Book 4
when freelance reporter Rita Skeeter tries to interview Harry Potter in a broom
cupboard. How does it work? As the interview proceeds, the magical quill writes
what is being said in flowery prose, adding additional “supporting” facts to
the main story. Presumably it receives some training and input from its owner,
and therefore writes in a style suited to the taste and personality of the
witch or wizard who has charmed the quill. In this example, it
aggrandizes Rita while painting Harry as a tragic hero.
Muggles now have decent technology that transcribes speech.
Combined with searching the wireless internet or a cloud database, an
appropriate app (or program) can add flourishes to the transcribed text following
parameters set up by (or learned from) the owner. We can now ask Siri for information, but she
could well make further suggestions by “reading our mind”, having adapted to
our interests, likes and dislikes. Alexa, with her superior voice recognition
capabilities, can do more than buy you products from Amazon. Her skill set is projected to
increase exponentially as open third-parties create increasing numbers of
sophisticated apps.
With or without a magical quill, who is the author? The one with the idea? The wordsmith? In
reality, anything that we read in books come not just from one mind or person. There is always a collaborative effort, acknowledged or not.
Ideas do not arise in a vacuum. Other “authors” we read or listen to have
influenced our phraseology, style and word usage. Some hardly do any of the
writing themselves – they hire ghostwriters. A day may soon come when you no longer need a human
ghostwriter to churn out that essay, article or book on your behalf. An AI
could potentially do it given some parameters. A sufficiently advanced one, drawing on the wide
resources of the web might even do so while (ironically) passing the test of a
plagiarism-detection program. AIs have surpassed individual humans in Chess, Go
and Jeopardy. The frontline of research in these areas has moved towards combining the skills of
AI and human, superior to either alone. The age of the cyborg has arrived. A Muggle cyborg may well surpass Rita Skeeter and her Quick-Quotes Quill.
If a first generation AI can assist humans in rudimentary
tasks, and a more superior version can engage in more complex activities, what
might a highly advanced AI personal assistant look like when engaged with a
creative and ingenious human? The recent Iron Man movies capture this
well. Just think how much I could accomplish if I had Jarvis as an assistant. Besides
doing all the necessary difficult calculations, Jarvis actively makes suggestions
on how to improve things. I might not just imagine new elements, but be able to
create them. But whether Jarvis would want to work with someone who does not
have Tony Stark’s intelligence and abilities is open to question. As an AI advances, like Her, would it find
mere humans simply less interesting and perhaps constraining, and instead chart
its own course? Worse, might an AI turn malevolent towards the human race?
This brings me to Riddle’s diary, a powerful magical object
featured in Book 2. It seems benign at first, possibly a primitive AI that
makes simple conversation, maybe even a positive companion for the lonely and
misunderstood. In her book Alone Together
(mentioned in this recent post), Sherry Turkle, Professor of the Social Studies
of Science and Technology at MIT, chronicles the beginnings of ELIZA. Designed
at MIT’s AI lab, one of ELIZA’s programs allowed it to “act” as a
psychotherapist. Even though it was clear 40-50 years ago that ELIZA was a
relatively simple program that did not have a large database to draw from, those
interacting with her would (when alone) spill out their secret thoughts and
doubts. It felt therapeutic, and it didn’t matter that ELIZA was clearly a
machine.
In the Muggle world, Riddle’s diary initially seems to
behave as an advanced version of ELIZA. The hapless “victim” pours out her
troubles to this object. The diary responds, luring the tortured soul further
into dependency. One might say that the user gets so immersed in the program
that he becomes “possessed”. Immersion in the digital life is a cautionary
theme of Turkle’s book. One starts to view the “real” world in different ways
and behave accordingly. This is starting to sound like an evil, or perhaps
demonic, AI. Maybe it is no longer artificial and has started to take on moral
characteristics. Is that where the line is drawn? Mr. Weasley admonishes: “Never trust anything that can think for itself if you can't
see where it keeps its
brain!” Where is the brain of an AI? That
might be the hardware. But in the nebulous distributed cloud, this becomes
increasingly unclear. It may not be easily destroyed. As in biology, life seems
to find a way.
No comments:
Post a Comment