June 10th, 2014
|…that’s not Turing’s Test…
“Turing Test” image from xkcd. Used by permission.
There has been a flurry of interest in the Turing Test in the last few days, precipitated by a claim that (at last!) a program has passed the Test. The program in question is called “Eugene Goostman” and the claim is promulgated by Kevin Warwick, a professor of cybernetics at the University of Reading and organizer of a recent chatbot competition there.
The Turing Test is a topic that I have a deep interest in (see this, and this, and this, and this, and, most recently, this), so I thought to give my view on Professor Warwick’s claim “We are therefore proud to declare that Alan Turing’s Test was passed for the first time on Saturday.” The main points are these. The Turing Test was not passed on Saturday, and “Eugene Goostman” seems to perform qualitatively about as poorly as many other chatbots in emulating human verbal behavior. In summary: There’s nothing new here; move along.
First, the Turing Test that Turing had in mind was a criterion of indistinguishability in verbal performance between human and computer in an open-ended wide-ranging interaction. In order for the Test to be passed, judges had to perform no better than chance in unmasking the computer. But in the recent event, the interactions were quite time-limited (only five minutes) and in any case, the purported Turing-Test-passing program was identified correctly more often than not by the judges (almost 70% of the time in fact). That’s not Turing’s test.
Update June 17, 2014: The time limitation was even worse than I thought. According to my colleague Luke Hunsberger, computer science professor at Vassar College, who was a judge in this event, the five minute time limit was for two simultaneous interactions. Further, there were often substantial response delays in the system. In total, he estimated that a judge might average only four or five rounds of chat with each interlocutor. I’ve argued before that a grossly time-limited Turing Test is no Turing Test at all.
Sometimes, people trot out the prediction from Turing’s seminal 1950 Mind article that “I believe that in about fifty years’ time it will be possible to programme computers, with a storage capacity of about \(10^9\), to make them play the imitation game so well that an average interrogator will not have more than 70 per cent. chance of making the right identification after five minutes of questioning.” As I explain in my book on the Test:
The first thing to note about the prediction is that it is not a prediction about the Test per se: Turing expects 70 percent prediction accuracy, not the more difficult 50 percent expected by chance, and this after only a limited conversation of five minutes. He is therefore predicting passage of a test much simpler than the Test.
Not only does the prediction not presuppose a full Turing Test, but it could well be argued that it had already come to pass with the advent of Joseph Weizenbaum’s Eliza some thirty-five years early. Weizenbaum developed a simple computer program that attempted to imitate the parroting conversational style of a Rogerian psychotherapist…. Although the methods used were quite simple – repeating the user’s question after adjusting some pronouns, throwing in a stock phrase every now and then – the result was, in its own way, extremely convincing.
Second, “Eugene Goostman” uses the very techniques that began with Weizenbaum’s “Eliza” program from the 1960’s. We see the same tricks – repeating the judge’s statements with simple substitutions, keyword-triggered responses, falling back on vague or unresponsive replies, and the like. Those tricks are no more successful than they have been in the two decades of runnings of the Loebner Prize Competition, another ill-conceived attempt at running a Turing-like test. And there too, entrants used the trick of having their programs emulate humans with built in excuses. “Eugene Goostman” purports to be a non-English-fluent child, rather than a fluent adult.1
In fact, anyone with the slightest familiarity with chatbots, the modern incarnations of “Eliza”, could unmask “Eugene Goostman” quickly. (The judges at the recent competition were apparently not so familiar.) Scott Aaronson has provided a transcript of his own interaction with “Eugene Goostman”, which shows the telltale signs as soon as the computer’s first reply, which blatantly ignores an utterly trivial question and tries to change the subject:
Scott: Which is bigger, a shoebox or Mount Everest?
Eugene: I can’t make a choice right now. I should think it out later. And I forgot to ask you where you are from…
Here’s an appeal to cute, vague language:
Scott: How many legs does a camel have?
Eugene: Something between 2 and 4. Maybe, three? :-))) By the way, I still don’t know your specialty – or, possibly, I’ve missed it?
And here’s the repetition of a canned response:
Scott: No, I need to know that you’re not a chatbot. Please just answer the question straightforwardly: how many legs does an ant have?
Eugene: Something between 2 and 4. Maybe, three? :-))) Oh, what a fruitful conversation;-)
Here’s a canned keyword-triggered response, nonresponsive as usual:
Scott: In your opinion, does your existence demonstrate any idea or principle that wasn’t demonstrated just as convincingly by ELIZA, Joseph Weizenbaum’s chatbot from the 1960s?
Eugene: Eliza was a break-thru. All the bots after it were nothing but weak parodies, claiming to have “revolutionary improvements”.
And here’s “Eliza”‘s trick of rewriting the judge’s utterance after pronoun replacement and (poorly executed) synonym substitution:
Scott: OK, I’ll count that as the second sensible thing you’ve said (grading generously). Don’t you think that a more convincing chatbot could be created, by using Google to trawl the entire web for plausible-sounding answers (analogous to what IBM’s Watson did)?
Eugene: Yes, I do think that a more convincing chatbot can be built by using google to trawl the all web for plausible – sounding – I hope I have right to have my own opinion. Maybe, let’s talk about something else? What would you like to discuss?
Literally every one of “Eugene”‘s responses reflects its “Eliza”-like programming. It would be amusing, if it weren’t so predictable.
In summary, “Eugene Goostman” is not qualitatively superior to other chatbots, and certainly has not passed a true Turing Test. It isn’t even close.
- In a parody of this approach, the late John McCarthy, professor of computer science at Stanford University and inventor of the term “artifical intelligence”, wrote a letter to the editor responding to a publication about an “Eliza”-like program that claimed to emulate a paranoid psychiatric patient. He presented his own experiments that I described in my Turing Test book: “He had designed an even better program, which passed the same test. His also had the virtue of being a very inexpensive program, in these times of tight money. In fact you didn’t even need a computer for it. All you needed was an electric typewriter. His program modeled infantile autism. And the transcripts – you type in your questions, and the thing just sits there and hums – cannot be distinguished by experts from transcripts of real conversations with infantile autistic patients.”↩
June 4th, 2014
To university administrators and librarians:
|…enablement becomes transformation…
“Shelf of journals” image from Flickr user University of Illinois Library. Used by permission.
As a university administrator or librarian, you may see the future in open-access journal publishing and may be motivated to help bring that future about.1 I would urge you to establish or maintain an open-access fund to underwrite publication fees for open-access journals, but to do so in a way that follows the principles that underlie the Compact for Open-Access Publishing Equity (COPE). Those principles are two:
Principle 1: Our goal should be to establish an environment in which publishers are enabled2 to change their business model from the unsustainable closed access model based on reader-side fees to a sustainable open access model based on author-side fees.
If publishers could and did switch to the open-access business model, in the long term the moneys saved in reader-side fees would more than cover the author-side fees, with open access added to boot.
But until a large proportion of the funded research comes with appropriately structured funds usable to pay author-side fees, publishers will find themselves in an environment that disincentivizes the move to the preferred business model. Only when the bulk of research comes with funds to pay author-side fees underwriting dissemination will publishers feel comfortable moving to that model. Principle 1 argues for a system where author-side fees for open-access journals should be largely underwritten on behalf of authors, just as the research libraries of the world currently underwrite reader-side fees on behalf of readers.3 But who should be on the hook to pay the author-side fees on behalf of the authors? That brings us to Principle 2.
Principle 2: Dissemination is an intrinsic part of the research process. Those that fund the research should be responsible for funding its dissemination.
Research funding agencies, not universities, should be funding author-side fees for research funded by their grants. There’s no reason for universities to take on that burden on their behalf.4 But universities should fund open-access publication fees for research that they fund themselves.
We don’t usually think of universities as research funders, but they are. They hire faculty to engage in certain core activities – teaching, service, and research – and their job performance and career advancement typically depends on all three. Sometimes researchers obtain outside funding for the research aspect of their professional lives, but where research is not funded from outside, it is still a central part of faculty members’ responsibilities. In those cases, where research is not funded by extramural funds, it is therefore being implicitly funded by the university itself. In some fields, the sciences in particular, outside funding is the norm; in others, the humanities and most social sciences, it is the exception. Regardless of the field, faculty research that is not funded from outside is university-funded research, and the university ought to be responsible for funding its dissemination as well.
The university can and should place conditions on funding that dissemination. In particular, it ought to require that if it is funding the dissemination, then that dissemination be open – free for others to read and build on – and that it be published in a venue that provides openness sustainably – a fully open-access journal rather than a hybrid subscription journal.
Organizing a university open-access fund consistent with these principles means that the university will, at present, fund few articles, for reasons detailed elsewhere. Don’t confuse slow uptake with low impact. The import of the fund is not to be measured by how many articles it makes open, but by how it contributes to the establishment of the enabling environment for the open-access business model. The enabling environment will have to grow substantially before enablement becomes transformation. It is no less important in the interim.
What about the opportunity cost of open-access funds? Couldn’t those funds be better used in our efforts to move to a more open scholarly communication system? Alternative uses of the funds are sometimes proposed, such as university libraries establishing and operating new open-access journals or paying membership fees to open-access publishers to reduce the author-side fees for their journals. But establishing new journals does nothing to reduce the need to subscribe to the old journals. It adds costs with no anticipation, even in the long term, of corresponding savings elsewhere. And paying membership fees to certain open-access publishers puts a finger on the scale so as to preemptively favor certain such publishers over others and to let funding agencies off the hook for their funding responsibilities. Such efforts should at best be funded after open-access funds are established to make good on universities’ responsibility to underwrite the dissemination of the research they’ve funded.
- It should go without saying that efforts to foster open-access journal publishing are completely consistent with, in fact aided by, fostering open access through self-deposit in open repositories (so-called “green open access”). I am a long and ardent supporter of such efforts myself, and urge you as university administrators and librarians to promote green open access as well. [Since it should go without saying, comments recapitulating that point will be deemed tangential and attended to accordingly.]↩
- I am indebted to Bernard Schutz of Max Planck Gesellschaft for his elegant phrasing of the issue in terms of the “enabling environment”.↩
- Furthermore, as I’ve argued elsewhere, disenfranchising readers through subscription fees is a more fundamental problem than disenfranchising authors through publication fees.↩
- In fact, by being willing to fund author-side fees for grant-funded articles, universities merely delay the day that funding agencies do their part by reducing the pressure from their fundees.↩