On artificial intelligence, museums, and authenticity

Cronut credit: https://www.flickr.com/photos/jpellgen/29762364771

Way back in the day…I had a weblog. Weblogs were places where the olds posted thoughts online and shared links to other web pages that often ended in “.html.”

One day in the early aughts, long before hashtags and cronuts, I wrote a post about authenticity and privacy in the digital age. My half-baked argument was that we couldn’t stop the frenzied land grab for personal data, but we did have control — even if tenuous — on the veracity of this newly valuable digital stuff. And isn’t authentic truth a foundation of value?

…Right? (…Right?)

Fast-forward to our current, post-cronut age. Fake profiles and fake user data are so entrenched in internet culture that there are startups offering to fix the problem and startups offering to supply the problem. Elsewhere, authenticity — as modeled through authentication — has evolved in awkward steps. The username & password method has long been the standard way to authenticate users and provide a personalized experience, but as breach after breach after breach have taught us, it really stinks as an architecture. Not surprisingly, the tech sector has been trying to move away from the username/password approach for some time.

Meanwhile, personalization continues to gain momentum through passive architectures like AI-assisted cameras. McDonald’s is soon using license plate data from drive-through cameras to upsell a McFlurry, because their machine learning algorithms know you probably want one, and Chinese airport kiosks use facial recognition AI to authenticate and give personalized flight info. Retailers use cameras for security purposes and increasingly for transaction and marketing reasons. We’re opting into the personalized/authenticated future simply by moving through the world, whether it’s in cars, airports, stores, or even…museums.

In a sense, authenticity and authentication is a negotiated space. If, for example, I exchange photogrammatic access to my face for access to your country, then this swap could be considered as an exchange of authenticity (I am who I say I am) for the ability to continue my travels. (At least, there’s an awareness of when this exchange took place, with a cruelly unflattering photo taken of my creased face and matted hair after a long flight.)

It seems like there’s a slow and tectonic shift going on here, and the cultural sector is in an interesting space…whether it wants to be or not.

In the cultural space, authenticity comes to life every time a visitor engages with an exhibit or object on display. In each wall label and supporting texts, the museum says: We’ve authenticated this object as this particular thing, created by this person or people, it happened around this date, and we gained this understanding as a result. And visitors, for the most part, accept these statements as a widely assumed aspect of a museum-going experience. (With some notable exceptions.)

So authenticity can be imagined as a negotiation of something that’s both given — as with a password, or derived from facial features — and as something received, as in accepting the authoritative statements in a wall label or exhibition catalogue. In this pretty limited definition it involves at least two parties in agreement over what’s true.

And here’s where a shift is taking place: As museums discover better ways of personalizing experiences to meet engagement goals — via AI, for example — and as visitors acclimate to public spaces as “opt-in” places, the tension between “visitor/patron” and “customer/user” will shift museums from being arbiters of authenticity increasingly to becoming consumers of it.

In this shift, what’s the obligation to disclose how a personalized experience is delivered, on a website or in a touch table, for example? Are museums held to a different standard than, say, shopping malls or airports? In the name of best practices and disclosure, museums already use provenance research to identify problematic histories and disclose real vs. replica objects in galleries. When AI is put to work in the name of mission-centered goals and authenticating human-centered data, should there be a corresponding disclosure of when/why/how it was used?

(Spoiler alert: Saying “yes” is easy but the reality is hard. Answering this in the affirmative puts GLAMs in a nearly impossible position with respect to the current “black box” algorithms associated with many third-party AI platforms. Yet I think — perhaps unrealistically — that GLAMs can lead by example and advance this cultural dialogue where it counts.)

Lately I’ve been thinking that any bright future involving AI should have three parts, on a roughly equal playing field:

1. Development of transparent, bias-free algorithms by the tech sector
2. Oversight of these algorithms through publicly supported legislation
3. Education of the public and extending ‘digital literacy’ to include AI

Items #1 and #2 are huge, complex issues. Yet item #3 is in the domain of the cultural heritage space. Given its enviable place with regard to authenticity, and the impact of AI across the entire topic in this era of “fake news,” GLAMs already have some skin in the game.

Sidebar: All that said, this isn’t meant to be a “get off my lawn” lament about the death of privacy and transparency in an information economy. Successfully personalized experiences help foster connection and they depend upon authentic data…and museums have a stake in the cultural idea of authenticity. Authentic, human-centered data and how AI leverages it (and reinvents it) is the economy getting built and negotiated — regardless of whether transparency exists or not.

If this isn’t a lament then, it’s a meandering attempt to reconsider authenticity in this age of mechanical reproduction. Museums are purveyors of authenticity and consumers of it, as more AI-driven marketing tech and CRM systems rely upon visitor data of the inferred and the self-identified, authentic kind. GLAMs invest in and derive considerable authority from the presentation of the authentic and can help drive the cultural dialogue. From Benjamin to Warhol to Obvious, GLAMs have been an arbiter of authenticity — and the value of it — in every era. Whether it’s welcome or not, the sector’s relationship to authenticity is changing, thanks in part to AI and emerging economies of truth.

This was think piece #5 in the “On artificial intelligence, museums, and hot dogs” series. Thanks for reading! Thoughts? I’m at @CuriousThirst.