Dialogue session between writers and the National Library re: Generative AI

I first posted the following on social media on 16 January, following a 14 January dialogue session at the National Library in response to our collective statement on AI. Sorry for posting it so late here!

Remember how, in our Collective Statement from Singapore’s Literary Community on AI, we asked NLB for a consultation about this technology?

Well, on Tuesday night, we had that consultation—sort of. Gene Tan, Chief Librarian & Chief Information Officer, had privately invited a bunch of writers, including myself, to a 6:30pm session in the Possibility Room. Terribly quick, but he *did* have a few extra days to plan this, since, according to him, *multiple* writers shared the statement w him before it went public. (I asked him who. He wouldn’t say.)

Now, I’m used to government institutions responding to activism through censure or denial. Gene’s chosen strategy, however, wasn’t just engagement: it was an all-out charm offensive. He catered dinner for us, allowed all to participate (this was publicized only on private platforms, so as not to attract the tech crowd), pleaded us to “learn how to exist between certainties” (a quote from Conclave) & emphasized that he sees himself as part of the lit community.

More bizarrely, he claimed this was a “date”, said that he was “heartbroken” when he saw the statement, made himself look vulnerable by asking us not to record his voice because he’s self-conscious, kept veering off on personal tangents, & pleaded w us to take a bookfie w him at the end to promote Amanda Lee Koe’s Sister Snake. (Everyone said no; we were too weirded out.) Also, he took up most of the night lecturing us on NLB’s perspective, leaving limited time for listening to our concerns—though he *did* let us speak, as you’ll see.

Basically, he doesn’t feel it’s fair for us to say NLB’s been uncritical endorser of AI. From his perspective, they only started doing programs when they could be sure it wasn’t just a passing trend, then took special pains to respect IP & factuality: they used source material that was uncopyrighted or expressly permitted by authors; they put notes about the ethics of AI alongside the opening pages; they severely limited AI creativity so the S Rajaratnam or 700 Years of Singapore ChatBook wouldn’t suddenly start hallucinating nonsense.

The much-maligned “Children Write: Publish a Book with Gen-AI” workshop by PicoLibo began w a 30 min intro to ethical frameworks. The SURE program’s been putting up info on all aspects of AI, including its controversial issues. & on top of all this, only 4% of all library programs in the past 2 years have been AI-focussed.

OK—great for clarification. But, as the audience pointed out, this isn’t really enough. The average NLB patron isn’t seeing the advisories; they’re just seeing the big flashy Playbrary displays & the titles of events, which are all either pro-AI or seeking balance—i.e. anti-AI arguments are always the footnote, never the focus. Their Tiktoks about controversies are always delivered by “AI enthusiasts” & the bulk of their education is about misinformation & deepfakes, not ethics. & some topics end up buried—the “AI and Sustainability Explained” YouTube they boasted of has a total of only 6 likes!!!

In the end, what NLB’s doing is normalising AI. Sure, they can install all these internal guardrails, but once a kid leaves a workshop or StoryGen showcase, they’re gonna use Chat GPT. The warnings really aren’t that explicit, so users end up being uncritical, even if NLB officially is.

Also, don’t give me that sh*t about, “The Beast has been released.” We don’t have to normalize what we don’t want. Anyway, this is a country that bans chewing gum. They’d start caning people for using AI if they felt like it.

Furthermore, as one attendee said, StoryGen isn’t really very good! She tried it out, & realized they’ve only created the same kind of story engine you could create w flipbooks or madlibs. Is it *actually* stimulating & educating kids, or is it distracting them w bright lights? (I wish we’d held a vote in the room about whether NLB should proceed w the 3rd iteration of StoryGen, cos a guy later emphatically told me no.)

Other points from the Q&A. Was NLB directed to do this cos of Smart Nation 2.0? (Gene says nope!) Shouldn’t we be talking to IMDA or MDDI instead? (I said NLB’s still the easiest dialogue partner we writers have, but sure, let’s take this further.) Shouldn’t NLB be doing more education on the history of AI, copyright, sense-making tools rather than sense-breaking tools? (Some of the answers got lost in the overload of questions at the end.)

To close—Gene said we should date more, i.e. that dialogues like this should happen again, & that his social media is open to the writing community to make ourselves heard. Which you *can* read as an effort to co-opt & silence, invalidating more public activism like this statement. But I also think he’s telling the truth. He wants to be trusted by the lit community. He wants us to get into his inbox.

But will policies change? I don’t think they will, unless we hold NLB accountable. & as much as Gene wants to make this personal, this *is* institutional & professional. Hell, it’s *national*, since Smart Nation 2.0 is a thing.

Big thanks to NLB for opening up to dialogue.

& a reminder to everyone else—this conversation isn’t yet closed.

More pics (including Gene’s slides) on my Instagram and Facebook.