Evans is a place on Earth: Tools You Can Trust:AI @ the Science Gallery.

"AI is a tool. The choice about how it gets deployed is ours" - Oren Etzioni, founding CEO of the Allen Institute for Artificial Intelligence

Wesley Goatey - Newly Forgotten Technologies (2023)

AI is not the future. AI is the present. As well as, probably, the future. But let's not split hairs. AI is here and it is having a huge effect on our lives right now. It monitors us, it analyses us, it categorises us, and it profiles us. It helps us access healthcare (or not), it helps us connect with people (or not), and it even helps us find love (or not).

But all the AI we have was, initially, programmed by human beings and human beings have preferences, biases (unconscious or conscious), and embedded ways of thinking and these, surely, have been transferred to the AI, and the algorithms, we now use to find out way through ordinary life. Surely?

Maybe the Science Gallery's free exhibition, AI:Who's Looking After Me?, would help me to understand a bit more about AI. But then - maybe not. It's not a good look to gripe about a free exhibition (though the extortionate price I paid for a cappuccino and a slice of Victoria sponge cake in the cafe afterwards more than made up for that) but I left the exhibition feeling somewhat disappointed. Maybe that will be my/our experience of AI in the end.

The questions the exhibit asks are valid ones. How are AI apps designed? What impact are they having on our lives? Are we all seen and valued in the same way? How much do we know about these systems and how much can we trust them? Once they're set up can we influence or change their behaviour?

The answers, however, were - at best - incomplete. That's understandable. There's still a lot we don't know about AI and how it will shape our future world. But, also, most of the contributions to the exhibition didn't really tell me anything, some of them looked dated already (only one of them intentionally so), and at least one of them wasn't working properly. Again, maybe that's how AI will be. Like all other software, a helpful tool and a massive frustration at the same time. A future as imagined by Charlie Kaufman.

James Bridle - Autonomous Trap 001 (2017)

The first work I saw was a pretty good example of just that. James Bridle has reimagned AI as a "tool for resistance and collaboration" and he's done so by travelling to Mount Parnassus in Greece (though the concept would have worked anywhere, he obviously just fancied a Greek holiday and why not?). The car in Bridle's work is surrounded by a salt circle - made by the car's driver. It's a self-driving car and the patterns made by the salt circle resemble lines that the car will interpret as "no entry" so the car cannot leave the circle.

It is trapped. The film only lasts about ninety seconds (much shorter than some of the less interesting fare on offer) but you get the point immediately. You can fool AI quite easily. The human brain is more creative, it seeks intuitive solutions, and it has a sense of humour (or at least most human brains do). All things that AI, so far, is lacking. Especially, judging by this exhibition, the sense of humour.

Sprout, by Air Giants, at least has a go. It's a soft, huggable, robot (I hugged it, it didn't hug back, typical AI) that changes colour and wobbles about as you interact with it. The idea is that Sprout (and who named it that?) is learning about how humans use their personal space as a form of non-verbal communication and then, eventually, facilitating more effective communication between people and robots.

Air Giants - Sprout (2023)

The 'soft robot' Sprout is very much a work in progress and at least it's an interesting idea. I was very disappointed by Fast Familiar's Looking for Love. Three desks, set up like an Internet Cafe in a Student Union and designed to be something like a tamagotchi. What could be more futurisic?

The idea is you take a seat (I did) and teach a robot about romance (I tried and failed, to be honest I don't know much about it myself)/ You're supposed to be (and don't split your sides at this) a "wing-(hu)man" to an emotionally illiterate computer who just wants to find love out there in the big old scary world.

So you have a 'conversation' with the computer that, to me, resembled a text adventure from a 1980s home computer game (anyone remember I Smell A Wumpus?, thought not). According to the makers, users will work with the computer to "distil the essence of romance from movies, poems. and song lyrics" but that's not what happened to me. It asked me a load of inane questions, mostly answering YES or NO, and then crashed.

I gave up and read the Post-It notes that seemed to be part of the installation. I had a very clear feeling that these Post-It notes were not submitted by visitors to the exhibition but by the makers of Looking for Love. That's because ordinary visitors would have written funnier, ruder, things. "Azad was here", "LOVE IS LOVE", and, best/worst of all, some lyrics from Eurodance legend Haddaway's 1993 number two hit 'What Is Love?".

Fast Familiar - Looking for Love (2023)

Oh, baby don't hurt me. Looking for Love inspired me, to use some appropriate Geordie vernacular, to hadaway and shite and go and look at something else. Something else, it turned out, that was equally disappointing. Munkhtulga Battogokh's What is Essence? I couldn't really see the point of this at all. In theory, it's a work that examines how humans and machines see and attribute meaning. In reality it's a blurred screen of computer, AI, generated images.

Some quite realistic looking, others less so. Then there's loads of guff about "lenticular print", "Stable Diffusion", and "Midjourney" which I suspect you have to work in, or have a more then passing interest in at least, AI to understand. The idea is that if you ask AI to draw, for example - their example, a "Mongolian person" the AI may only know stereotypes, possibly harmful ones, of what Mongolian people look like. How do we deal, this work asks, with this cultural imbalance?

Munkhtulga Battogtokh - What is Essence? (2023)

Salome Bazin - Sentient Beings (2023)

It's not a question the work even attempts to answer and, to be honest, after a couple of minutes looking at it you'll be past caring anyway because it's so dry and lacking in engagement. Still, at least you can interact with it (if you can be bothered). Salome Bazin's Sentient Beings doesn't offer you even that. In fact it expressly forbids you from touching, or interacting with, the work.

It is, though you'd be hard pushed to notice, an "immersive, evolving soundscape" and a "whispering gallery" that invites you to consider how AI is trained and programmed. It did nothing of the sort for me. I couldn't make head nor tail of it. It looked like an office desk that had been left empty after the attendees of a 'break out meeting' all got bored and decided to go outside for a vape instead.

I started to think that the biggest danger of AI is that it may take over by boring us to death. At least though, or so it seems, our pets will be fine once we're gone. Blast Theory's Cat Royale imagines three pet kitties called Clover, Ghostbuster, and Pumpkin (yes) living in a custom built environment and being looked after by AI.

Their feline utopia is full of ledges to relax, dens, and places to climb. A robot arms throws balls to the cats, dangles feathers, and offers snacks. Human observers can measure how successful the AI is in looking after the cats (so you can feel like those demotivated ticket inspectors at railway stations who have been replaced by machines but still have to be there for when the machines go - regularly - wrong) and if you've got a spare seven hours and ten minutes you too can be one of these human observers.

Blast Theory - Cat Royale (2023)

I like cats - but not that much. Certainly not ones that aren't even real and certainly not ones called Ghostbuster. A couple of minutes looking at Sarah Selby's Between the Lines felt like a lifetime caring for imaginary cats with stupid names. Even the description of the work is enough to make you glaze over:- "Between the Lines highlights the power of bureaucratic systems and the risk that using algorithms within UK immigration and justice services could further disconnect individuals from the decisions made about them".

Of course, immigration and how fair it is (currently, not very) is a very important subject and even something of a hot topic but I'm not sure Selby's installation, which looks like an abandoned office - yes another one, tells us anything at all about the subject it purports to. Though, on the plus side the pens that make up part of the work will be distributed to individuals working on the frontline of UK immigration so at least some good will come of it. Hopefully.

Sarah Selby - Between the Lines (2023)

I had higher hopes for Seo Hye Lee's How Loud is Too Loud? An interactive work that, unlike most of of the others on show, you can actually interact with. How exciting. In theory at least. Here's what happened. I stood underneath a speaker and noises came out of it. Throbbing noises, looping noises, drone noises. For the most part, quite quiet noises.

As instructed I responded by telling the AI if the noises were "too soft", "normal", or "too loud". They were all, bar one, "too soft". One was "normal" (and that was pushing it) and none were "too loud". The idea being that AI such as this could be used to help people with hearing problems (perhaps I have one if I felt most of the noises were "too soft", I think I just like loud music but maybe that has fucked my hearing) and that idea is a good one. The way it didn't work is that once I'd submitted my thoughts I got precisely zero feedback and simply wandered off to the next exhibit. Clearly Seo Hyo Lee has work still to do.

Seo Hyo Lee - How Loud is too Loud? (2023)

Mimi Onuoha - The Future Is Here! (2019)

Mimi Onuoha's The Future Is Here! is a presentation of the domestic sites where crowdsourced labourers working for companies like Sam's, Appen, and Clickworker (companies that deal with huge qualities of data to help speed up image recognition) toil away. It's not particularly illuminating but it does, at least, shine a light into the darker corners of the AI industry and look at the humans that make it all possible.

A for effort. Sofie Layton's Does AI Care? takes on an even more serious subject. It looks at 'care' and the experience of young adults living with cancer as well as how AI is now used for medical imaging, radiotherapy, and even surgery. The idea here is that you enter a hospital waiting room, there's some frankly unconvincing curtains with stuff - bromides - written on them but other than that you get the idea, and listen to an audio piece that combines reflections on cancer with reflections on AI and care. It's all generated by an AI chatbot and it lasts for twelve minutes. I couldn't hear a word of it so the whole thing was wasted on me. Perhaps I shouldn't attend exhibitions in their final week where, sometimes - not always, things seem to have gone a bit to shit.

Sofie Layton - Does AI Care? (2023)

TripleDotMakers - Heartificial Intelligence (2023)

TripleDotMakers must have spent all of five seconds coming up with the title of Heartificial Intelligence for their work. To be fair, I didn't spend much longer looking at it. Which feels a bit mean because the work is based on "a group of nine young people who have congenital heart conditions". The instructions here is that you the visitor, in this case - me, place your index finger on a sensor and watch a series of hearts pulse in sync with your own heartbeat and your body's rhythms become date input for generative AI.

I did this, I give most things a go, and - and you may be noticing a theme as regards this entire exhibition - absolutely nothing happened. While AI may well change my life, and almost certainly already is, it seems exhibitions about AI don't quite cut it. But then the penultimate work in the show proved to be the very best on offer. Maybe all was not lost.

Wesley Goatey's Newly Forgotten Technologies imagined a near future in which dozens of discarded phones, tablets, smart speakers, and listening devices li(v)e in a dump telling stories about how and why they were thrown away. They imagine a future where our relationship with AI has changed and one of the stories, there are several playing at the same time and they are not always easy to hear properly, tells of a fictional future event in which the government has banned AI from snooping on us and all AI devices have been abandoned.

Wesley Goatey - Newly Forgotten Technologies (2023)

Jessie Montgomery - O-Horizon (2023)

That's the sort of dysfunctional future I came to this show to be warned about. To be fair, there were plenty of examples of how AI may provide untold benefits for us in the future but, for the most part, those examples were a bit crap. The Newly Forgotten Technologies told a story that you could understand, and get into, immediately. Not have to wait over seven hours for like the imaginary cats.

There was one last piece in a room that I thought was part of the cafe. Jessie Montgomery's O-Horizon is inspired by the "forest floor" (really?) but it's basically a large black box (think the Kaaba at Mecca or the monolith in 2001:A Space Odyssey) surrounded by our old friends, the Post-It notes.

The visitor is invited, and this time I didn't engage, to write on a Post-It note what they would, if they could - in some kind of almost perfect world, use AI for. Some didn't read the brief properly (instead writing about the danger of AI being in charge of nuclear weapons) and others were more idealistic (ending homelessness - though in a vaguely perfect world hopefully that would have already ended).

One commentator, however, cut right to the chase and suggested we could use AI to create 'SEX ROBOTS'. At least I didn't feel that was a fake visitor. It felt very much like the sort of comment a real visitor would make. A real visitor who can't get a real sex partner. Though AI, as stated in the opening line of this piece, is a tool it seems there are still plenty of us out there who think with their tools before their brains. Maybe AI would at least move us on from this (although it's debatable if that would be a good or bad thing). The trouble is that, like most technology and most power in the world, it seems inevitable that it will fall into the hands of men who, not to put too fine a point on it, could best be described as 'tools'. It's not AI we have to fear, it's the people that control it.