Presented by Sree Sreenivasan, co-founder and CEO of Digimentors, the session explored the practical realities of artificial intelligence in the cultural sector. Key takeaways from the presentation and accompanying discussions focused on: embracing AI as an augmentation, navigating the backlash, and practical applications where AI brings immediate value to museums, such as accelerating the production of audio tours and visual descriptions for accessibility, extracting visitor data, and creative brainstorming.

Transcript

Sree Sreenivasan:

Hello everyone. Welcome to this AI session at the Future of Museums Summit. My name is Sree Sreenivasan and our topic is AI and Museums: Beyond the Hype and the Backlash. I’ve had some history working in museums including as Chief Digital Officer of the Metropolitan Museum of Art and with museums around the world, including the Louvre Abu Dhabi, the West Kowloon Cultural District, and other beautiful places where we all gather in person. Here, of course, we’re gathering virtually, but we have learned through the pandemic that there are ways in which the magic of in-person can be replicated to an extent with these online tools. And that’s something that a lot of us would not have believed was at all possible before the pandemic. You can see my contact information and I would love to keep in touch and hear from you if you want to follow up on any of the things that we are talking about.

So I’m going to use some of these slides to share where we are in the world of AI. And then obviously you’ll be chatting and I’ll look at the chat and answer questions there as well. So as we think about this, I wanted to do a sense of how I think about the world of museums and technology in general. And I have had the chance to do a lot of presentations, as you can see here on the screen, digital lessons from the Met Museum, the Metropolitan Museum of Art to @MetMuseum: Thoughts on Going Digital. And I really worked on this fun project with Cynthia Round, who was chief marketing officer at the Met when I was chief digital Officer. And we have a presentation we do together called Don’t Shoot the Jellyfish and Other Marketing, Digital, Social and Mobile Lessons from @MetMuseum. And you have a couple of QR codes there, so if you’re really enthusiastic, you can go and get those.

Otherwise you can contact me and happy to send you them. And I worked at the Met till 2016, which is a lifetime ago in technology, and so much has changed since then. Back then we were talking about social media and how do we make sure social media continues to enhance the role of the museum and connect people with the world? And social media still matters, but not in the way that everyone’s talking about a newer piece of technology or newer aware piece of technology. And that is AI. I love this comment and tweet from a game designer named Joanna who says, “You know what the biggest problem with pushing all things AI is? It’s in the wrong direction. I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do laundry and dishes.”

And I thought that captures very well how a lot of us feel about the world of AI and how it’s the wrong priorities are surfacing, and that’s something that we will talk about as well. My work is based around a couple of big concepts. One of them is this, that the scarcest resource of the 21st century is human attention. How do we get it, grow it, and keep it when there’s so much competing for our time and our wallets? And that’s something that I think about a lot. The other is what I call my theory of the two footprints. I believe that we all have two footprints, our physical footprint of everyone who knows and loves us and our digital footprint, which is everyone else in the world. And if you don’t raise your digital footprint, you’re leaving impact, influence, opportunities, money on the table, resources on the table.

And so I encourage you to think about ways in which you can raise your digital footprint. That’s sort of my life’s mission now. And AI has a big role to play in that because that can do a lot to raise that digital footprint. The question is how does it do it and when does it do it? And we have framed this conversation as being kind of the early days of AI, AI as a teenager and what does that mean for us as people who work in the museum space and the cultural space. I’m not in it full time anymore. So I think I bring a little bit of an outsider perspective and a lot of the things I’m going to be talking about and sharing are in other fields, but reflect back on what’s happening in the cultural space. Here’s how long I’ve been teaching technology. Computers used to look like that, and I used to look like that. I did have hair once.

And what it tells you is that the work environment has also changed and how consumers interact with technology has also changed over that time. When I was at the Met, one of my former students had this wonderful line he wrote in an email. “We do some jobs for money and some for the honor and privilege of doing the work. And if we get to do both, we’re lucky.” I love that so much. I put it on the back of my Met business card, and I know many of you feel that way in your work. And that’s at the heart of what we want to be able to continue to do, the human part of what we do, the soul of what we do, even though technology is all around us. And so as I’m talking, I am keeping that in the back of my mind. How do we make sure that we keep that human side, the creative side of everything that we’re talking about?

And when I was at the Met, people would say to me, “Sree, how come you don’t have QR codes on paintings or next to every painting?” And I’d say, “Americans hate QR codes, they’re just not good at it and they don’t want it.” Well, now everybody uses QR codes because the pandemic launched the use of AI, sorry, of QR codes with menus. And so when it did that, then suddenly people were forced to use it. Well, the fact is that Apple put in a QR code reader in the iPhone in 2014, but didn’t tell anybody about it properly. And so you have the greatest marketing machine of all time doing a bad job marketing a great feature. And so there’s a lesson for all of us in museums that if we have a piece of technology or any project that we’re doing, what is our internal marketing plan, our external marketing plan, and you have to do it by specific things that you want people to understand.

You can’t just roll it out with a whole bunch of things. And that’s one of the lessons I learned in dealing with just even these QR codes and how did Apple not do what it should have done even better. Maybe the way we would’ve used QR codes in museums, in cultural spaces would’ve been different if they had rolled it out differently. And that’s something that will always be something we think about. Now, you recognize all of these channels. I put my email address in there because email still matters. In a TikTok world, email still matters. And as we think about tech, I like to go back in time and think about how tech has evolved and how hard it is to keep up with tech. Back in 1989, if I can take you there, Tim Berners-Lee, a young man working in Switzerland, sent his boss a proposal. And the proposal came back from the boss with just three words, “Vague but exciting.”

And the thing he was proposing was in fact, the World Wide Web and the World Wide Web is from which everything you see in front of you on the screen is derived. And the boss wrote back vague but exciting. And so it’s hard to know when you’re looking at some technology, is this useful? Is this something we need or is this something we should push aside? We don’t know that when you’re looking at technology. And keeping track of tech is hard, look at these. These are just the headlines from one day in the Bloomberg terminal of Apple headlines. That’s how much news is being generated about tech. And there are so many tools out there and brand names and things that are coming at you so fast.

One of the things I’d love for you to do is as I’m talking, if you have a piece of paper or a computer or phone, make little notes of things you want to try now, things you want to try later because it’s so hard to keep up with all this tech jargon and stuff. So for example, here’s a list of some brands you’ve never heard of them most likely, but they changed their names and became household names. So I’ll give you a minute to just kind of look at this list. And are there brands that you can recognize out of this? They all became very famous. I’ll give you two of them. The Sound of Music became Best Buy Electronics, the electronics stores around the country. And Auction Web became eBay. So let’s see if you can guess any of the others. I’ll give you a moment to just kind of think. 1, 2, 3, just kind of think through this and I’ll show you the list right now.

And you can see Backrub became Google, Cadabra became Amazon, Peekaboo, Snapchat, Burbn, Instagram, Matchbox, Tinder. I think that’s a really… Now you’re going to think, “That makes sense.” And Kibble became Netflix. But I’m showing you this because in the world of AI, we’re back at that point where we’re exposed every day to these kinds of names, and they may or may not mean anything to you. And some of them will go on, in fact to be world-changing tools and systems, and then in other cases they won’t. Well, I remember when I got to the Met, someone said the word tessitura to me, and those of you who work at museums absolutely know what it is. I was from outside the museum world, so I didn’t know what that is. I had to write it down. I didn’t even know how to spell it. And then it became a word that everyone in museums knows, right?

So that’s how technology evolves, and that’s something for us to think about as we deal with all kinds of tech in front of us. But we are at that moment now with AI where so much is going to change and these stats, even if they’re exaggerated, to say that we’re going to see a growth of seven to 15% to the global economy, this boost that’s going to be given to the global economies is incredible. Even if it’s exaggerated twice over. It doesn’t matter. It just tells you how important it is. But there’s also the other side, and here you’re looking at the exposure to AI-driven automation, meaning jobs can be lost. And you see Japan and the US about 25%, all the way to India at about 10%. Again, cut these in half, but in the US we’re looking at tens of millions of jobs being lost.

And I don’t think that… And by the way, many of them would also be in the museum space, in the cultural space, but I don’t think museum leaders are talking about it. I don’t think enough politicians are talking about it, thinking about it, bureaucrats. This is just not on people’s… They talk about AI, but the practical impact of it is not something that we have seen or understood in a proper way. So I urge you as museum leaders to bring up this topic and see what happens. We have this idea sometimes that the museum world is kind of isolated or insulated from the ravages of everyday life. When in fact lots of jobs that we have inside the museum space are jobs that exist outside. And therefore we should be constantly looking at the future and say, what does this mean for the jobs inside the museum when you see and hear about jobs outside the museum?

So a lot of us worry about jobs and as we do, I thought that the best explanation of how to think about it comes from Daniel Pink. And Daniel is a great writer about technology, and he said, “AI won’t replace humans, but humans using AI will replace humans not using AI.” And that I think is the right answer. That we all have to learn how to use this technology in order to survive and thrive in the new age with everything we see around us. So I’m showing you one of my favorite memes. This is Bart Simpson complaining about how it’s the hottest summer of his life. And his father says, “You mean it’s the coldest summer of the rest of your life.” And of course, I turn it into a ChatGPT joke where ChatGPT says, where he says it’s really scary. And his father says, “Remember, it’s the least scary, least powerful, least ubiquitous, least useful it will ever be.”

And that kind of frames where we are today as you think about all of this tech. And I promised you a lot of tools and things I’ll send your way. So here’s one called Misgif.app. And I have, as you see on the bottom right, the Made with AI logo that I’ve put on there to let you know that I’ve done that. And you can see it’s the ugliest Sean Connery has ever looked. There’s James Bond, and here’s me as Harrison Ford. And there I am again as Daniel Craig, but I also am here as Barbie. And when you see fun tools like this, you may say, “Well, we’ve had these tools. Face swap apps have always existed.” But what you’re seeing here is how much more realistic they are. And a lot of these tools allow you to keep your data and the tools in the work you’re creating on your computer rather than on server.

That’s something you should definitely look for as you play around with AI. Here’s another Made with AI. I asked for a picture of me and Elvis Presley, and I don’t think Elvis has ever looked so positively, but here is… They got the picture wrong. And then here you see it says that they got his written description right, but they got me completely wrong. They said I was chief executive, chief editor, and chief of the New York Times as well as Bill Clinton’s main speech writer, chief speech writer. And while my late mother would’ve loved that, that is not true. And we call this, hallucinations, when AI makes mistakes. And I’m on a one-man mission to eliminate that phrase and call them mistakes, errors, big effing deals. These are real things that do great damage in the world, and that’s something that we should be watching out for when we’re creating things with AI.

But AI isn’t new. Here I am in 2012 experiencing it when I wrote to my friend, Zeba, and it turned her into a Zeba or Grammarly, which many of you have been using for so long. But AI also isn’t new in museums. Here’s a look at in 2019, looking back at how AI has been used in museums, going back to 2010 from a very good paper by Elena Villaspesa and Ariana French. Elena is a former colleague at the Met. And it shows how you can think about all the tools that people are hyping up. And when you hear them, you can kind of think, “Which one makes sense for me and my museum.” This was the same thing we would say when you hear it about things with social media, same thing. What makes sense for me and my museum? That’s the thing that I think you should always be thinking about. When you get the slides, you can click on this and go to their paper and think about this there.

But the world of course changed in October 2022 when ChatGPT came around and all kinds of headlines. You’re seeing a four-week-old startup raised 105 million Euro. Lawyer cited six fake cases, a German newspaper fired a bunch of people, all because of ChatGPT and what it could do. This idea of generative AI. We saw Amnesty International, another AI… Initial AI criticized for using AI generated images. And we’ve seen particular museums and other places also get caught with using AI when they shouldn’t have. And so the machines can do whatever they want, but humans are making those decisions, and that’s an important part of the things that we should be thinking about. Here’s Cosmo Magazine, the world’s first AI cover. And then I put that Made with AI here and said, “It only took 20 seconds to make.” As you can see at the bottom of it.

This doesn’t mean that designers are out of work. It just means designers need to know how to use AI in order to be able to do their work and stay relevant in what we’re doing. And here are some mistakes that museums are making and can make with AI. Things like inaccurate or biased representations in AI-generated content, over reliance on AI for curation and interpretation. There’ll be a lot of opportunities for inaccuracies, stereotypes, biases that can come through. Privacy concerns with visitor data collection because AI will be used to mine that data. Accessibility issues with AI-powered interfaces. We saw the same thing with social media when it came around the web before it became Web 2.0 and before social media was much more accessible and a lot of good folks worked very hard to make social media accessible and we have to worry about that now as well.

And then job displacement due to automation as we talked about and misleading or sensationalized AI-generated art or exhibits. This is things that we are seeing. Now I want to reveal, as you can see with the Made with AI sign, that those three slides are all from talking to AI about mistakes that museums are making. And partly why I’m showing you that is that as I did my research and I thought about what’s out there, I could see that these are among the things that people care about, but we need to have an understanding among all of us that whenever we use AI for anything, we need to reveal that. A major newspaper announced that they’ll be AI-free. They will not use AI for anything to win their readers over, but how can they promise that? And it makes it sound like using AI is bad. It’s how you use AI and what you do with it that’s bad and problematic.

That’s something that I don’t think is clear to everyone. We want to think about the three waves of AI evolution and thinking about how it affects us in the museum world. So Mustafa Suleyman, who is the CEO of Microsoft AI and formerly at Google DeepMind, he talks about three kinds of AI, three steps of AI. Classification, generative, and interactive. Classification is where you’re teaching the computer, “That’s a tree, that’s a human, that’s a lamp.” The generative is where we are now, where you’re typing in something and something brand new is born. And then the third evolution that’s coming is where the AI is capable of communicating and operating autonomously. And to some of you that’s exciting and to others it’s really scary.

So if you use Zoom in your, you have seen this new thing called the AI Companion, you see it at the bottom of your screen and you see it to the right. So you can come in and say, “What are the action items?” Or “Catch me up.” And it’ll tell you what was said while you are away. That’s the second wave or the generative wave. And the interactive wave, as you can see in the current wave, somebody still has to do the work. It says, “Macy will send out meeting notes, Antoine will connect with Jada to…” Somebody has to do that work. All the AI is doing is collecting the information and telling you this is what’s going to happen. Well, in the next generation of this stuff, AI will just go do the report, send out the messages, do the work for you. And that’s where we in the museum world aren’t ready because we’re barely able to cope with what’s happening in this particular phase.

So something to think about as we go ahead and see all the new stuff that’s constantly coming at you. But really when people use AI to just talk to the machines and get information, I like this quote from another former student of mine. “The power of AI isn’t that it gives you answers. It’s that it forces you to ask better questions.” And so we talked about how AI is still in its very early stages, still very young. So one of the things we all want to do is to write better prompts, practice writing the prompts. It’s going to be a skill that will be on resumes, it’ll be on LinkedIn job descriptions. Are you a good prompt writer? So here are some ideas around writing prompts. But just like with children and adolescents, what you want is to ask it this question, how does it know? Why do you think that? What evidence do you have for that?

Any chance you’ve made mistakes? It’s like talking to a young person who does the math homework and then telling them, “Show me the steps.” And that’s what you want to do with AI. Don’t let it just go out there and do the work for you. Make it show you the steps. It’ll make a big difference in your work inside the museum world if you ask that question every time you’re out there using these tools. As with all technology, we need guardrails. We need to be thinking about how we can implement this in our museums, in our cultural spaces. And I love these six principles, transparency, inclusion, accountability, impartiality, reliability, security, and privacy. And you’ll look at this and say, “These work pretty well. This must be from Microsoft or something or something is the key because Microsoft was involved, IBM was involved.”

But this really came out of the Vatican, in something called RomeCall.org where you can go in and see how technology intersects with ethics and religion and things like that. And there are people from various religions involved, but the most important thing is to look at that date, February 2020. That’s how long back they were doing this before ChatGPT and all the buzz around this. So they’ve done a really good job of thinking about these things. And I think as museum leaders, we need to have ideas on how this works and what it means. I do some work with university presidents around AI because I think every university president needs to have a position on AI and how it’s being used inside their museum, what it means for employees, visitors for their school groups to larger issues around what it means for society. You could be interviewed about this topic at any time.

And so I think the more we can think about these issues and have our own ideas on them, the more successful we’re going to be. Now, the leaders of all these AI companies, many of them got together and issued a report about the dangers of AI, and they said it’s at the scale and danger of pandemics and nuclear war. And so you think it’ll be accompanied by an 800-page report saying what you can do. Well, that was the report, the entire thing. So they are acknowledging that they are creating a problem, but they’re saying, “Good luck everybody.” So as people in the cultural sector, we have an ethical quandary. Do we promote this? What does this mean? Where does it go and what is it doing to the planet? Because there’s a whole other side of this, what you’re seeing on the screen here.

That we need $1 trillion in funding for renewable energy to reverse the increased demands from these data centers of AI. And the more we use it, the more damage we’re doing to the planet. And museums have recognized that they have a role to play in all of this as leaders, as consumers, and something that we should be thinking about. Let’s talk a little bit about video and translations. When I worked at the Met, we spent a lot of time thinking about how we reach the world and how do we translate our content. We should pay a lot of money to do that. But some things we realized could be done with Google Translate, not the high-end scholarly work, but the things that would reach the masses. And now video plays a role in that. So here you’re going to watch as the UN Secretary General speaks in Hindi, even though he doesn’t speak Hindi. [foreign language 00:26:14].

You can see him talk about AI and it just won’t speak Hindi, you can see that that lip sync is pretty good. And so the question comes, should we be using this for all our work in terms of reaching people? And the answer was, no, we shouldn’t. Just because this video, this platform exists and it can do some things right, the importance of getting things right rather than just getting it out there is something we have to think about. There’s also an ethical question. If you have your museum director suddenly speaking 50 languages and he or she doesn’t speak those languages. When they go to that country and the first time they’re asked a question, then there’s an issue with that. So a lot of these questions around AI are about how you deploy it and how you use it in the work that you’re doing.

There are a lot of new cool video things that are coming, including something called Sora from OpenAI, which people are now testing. And it makes these beautiful video things that come from just your description. So you write a description and something pops out that looks real and people are very excited about it. But the fact is that you don’t know where this is going and what could happen with this. So I saw this tweet by someone who said, “People who think Sora will soon make movies are the same people who thought NFTs were art.” And you remember that debate within the art world about what the role NFTs play in art. And so this is taking a shot at those NFT folks, but we know that there’s a lot that can be done with art and AI.

Here is DALL-E, for my beagle’s Instagram, I asked it to make a painting of a Beagle eating the dosa, which is a South Indian crepe. And it did it pretty well and pretty fast. Here is an impressionist oil painting of someone teaching an AI workshop. Here’s a cyberpunk version of the same thing. And the interesting part is how diverse some of the images are. In some places the images are not diverse, but in this case, they are. And that’s something that we want to keep in mind as you’re looking at all of these tools. And here I’m going to give you a cheat sheet that you can use. This is the way to make art, images, or generate art on any art generator. You need an adjective, a noun, a verb, and a style of art. Do you want a pencil drawing? Do you want a photograph? Do you want Frida Kahlo style? All of those things you can do. And so this gives you a sense of how these tools work.

I’m going to show you an example of why the work we’ve done for generations in museums still matter. So here I’m showing you two images that came out of work that I did with a UN agency that works in East Africa where there are a lot of deserts and a lot of migrants crossing. And so here you see a picture of a motorcycle, and here you see a picture of a motorcycle. One of these has a lot of mistakes in it, and you can guess which one it is. You can tell which one it is. Of course it’s this one where you have the backward man. And what I tell folks is that there are mistakes here, but there’s also mistakes in here. But the mistakes here would only be caught by a motorcycle expert. Wouldn’t it be great if all mistakes were like this, so obvious? No, the mistakes are also going to be like this one.

And what we bring to the table in an AI world is interpreting, understanding, analyzing and explaining AI and catching it out when there are problems, and that’s why I’m confident that expertise is going to matter. As we look at the future, expertise and our background, our experience, all of that’s going to help us keep our jobs inside the world of AI and everything that it brings us. And there are things happening with audio as well. And audio tours are so big in museums as we know, and people interact very well with audio. And here’s something from a tool called elevenlabs.com. It says, “Give me a minute of your audio and I will clone it to do anything or say anything you want.” So I took the tool, I put in one of my favorite songs by Run DMC and here is some of the worst rapping you have ever heard.

I am Sree Sreenivasan and I’m here to say my Adidas walk through concert doors and roam all over Coliseum floors. I stepped on stage at Live Aid. All the people gave and the poor got paid. So apart from that terrible rapping, the fact is I only gave them 15 seconds of my audio instead of a minute. And what they’re saying is, “Give us a minute of your museum leader’s audio, and then you can have your museum leader do an unlimited number of audio tours or anything else that you want, that cloning of the voice.” And we’re going to see that in different parts of our lives. So many tools are going to be using this. So many use cases, including with the elections that are just a few days ahead, you’re going to see that people are going to get robocalls at scale of millions of messages going out saying, “We’ve already won the election. Stay home. You don’t need to worry, you don’t need to vote.”

And some people, when they’ve heard me talk about this, “Sree, you’re exaggerating.” Well, it already happened in the New Hampshire primary where somebody did this and he got caught and he was fined $6 million for sending out audio Joe Biden and saying, “I’ve already won the New Hampshire primary stay home.” So that’s what we’re going to see. So that interaction of what happens inside the museum and outside the museum is very important. And our awareness of following the news on these things makes a difference. And of course you heard that a lot of TV news programs, they say every family should have a technology safe word so that if you get a call from your mother or from your daughter or your father or son saying, “I need money.” You don’t get caught or you don’t get fooled like a congressman did who sent lots of money, thousands of dollars to somebody who thought was his daughter.

And those are all problems happening in real time now. So as we start to wrap up here, I wanted to show you how AI can think about what’s happening in the world around us. So here I use a tool, Perplexity, that some of you might know it’s a competitor to ChatGPT. And I said, “Suggest some slides for a presentation on AI and museums beyond the hype and the backlash.” And I did that after we made the present… That I did most of the presentation. So ideas are not from there that I used, but you can see here it says, “Here’s the introduction, what you should talk about. Here are some potential benefits for AI and museums, AI applications in museums, challenges,” et cetera. And it talks about all of this. And the conclusion is that AI offers transformative potential from museums, success requires thoughtful ethical implementation. And that’s pretty good.

Well, I went in and said, “Now do it in the style of Sree Sreenivasan and let’s see what they say.” So here now is the same topic that the AI was doing, but now in my voice, and one of the things because I’ve been a professor for a long time, you can see it ends with questions rather than conclusions. And it also has quotes as you can see here. “AI will quickly…” At slide number nine is a quote that says, “AI will quickly create a world of the haves and have-nots. I hope the museum sector will find itself on the right side of the equation.” So you can see even the style of presentation, it was able to change to be more like me because I’ve given a lot of talks and your leaders have given a lot of talks. Many of you have given a lot of talks. It’s learning all the time for better or for worse.

I want to conclude by showing you a couple of fun things and then some takeaways. A fun thing is the angry email translator. Museums still work on email as we know, and a lot of email gets written. So here is the angry email translator. You write what’s on your mind and it’ll make it soft and gentle and nice. That’s kind of fun though. Somebody said, “I want the opposite. I want it to go in reverse where you write something nice and it turns it nasty.” One of the big issues inside museums is the issue of bias. And that’s also there in AI. A lot of tools say, “Well improve your photograph, give us your photograph, we’ll make it better.” And here, the lady on the left gave her photograph and look what AI said is now a better photo of you. It changed her skin color, her eyes, her hair, her body, all of that.

It did because of the bias that’s built in to the technology. So as museum folks who bring ethics, big picture ideas, thoughtfulness to a lot of the work or all the work we do, of course. We still have to know that the tools we’re dealing with are made not by people who are necessarily sharing our values or our ideas or our thoughtfulness. And that’s something that we have to keep front and center as we’re working on everything to do with technology. It’s like in the early days of taking data and creating our customer relationship management tools. Those tools are meant to help companies harvest the data and squeeze profits. And while there may be people in the museum world with those ideas, what we want to do is serve a higher purpose. And so those tools, if you just leave them to themselves, they will not be like that and that will be a problem for what we do.

So now I wanted to share with you some takeaways about how we think about AI. Number one is that no technology can live up to the hype that’s there around the world. Every industry will be affected and lots of money will be made and lost and most careers will be upended in unexpected ways. And this is something that we’re all gathered here. You’re at a Future of Museum Summit, so you are the folks who are thinking about that. But how do we bring along our colleagues who are often just so happy and content in working in the things they’re already doing because they’re so specialized that they can understand the value and the opportunity and the dangers of AI. We want you to pay attention and learn new skills, that Daniel Pink quote that you’ve heard me say already and also that these are not hallucinations.

And then to be vigilant about the dangers. There was all this work that so many people did in the web and museums, social media and museums, blogs and museums. And we saw that every problem that was generated was generated by the companies and they did not police themselves. And that’s the same thing that’s going to happen with AI. AI companies will not fix the AI problems. And then finally, I believe that AI in the battle of machine and humans in the world of AI, that team human will prevail. But the question is what kind of humans will prevail? Will it be the same thing we’ve seen in all other technology where a few people in Silicon Valley, the tech bros get the money, the power, the influence, the impact, or will it be more spread out? And to do that, we need to step up and we need to get involved and we need to participate.

And only then will we have as a cultural sector, as the museum sector have a voice in this and change the way AI is headed, where it’s headed, the way it’s headed, and making sure that we can keep the work that we’re doing successful for centuries to come. Hello everybody. Thank you for watching. And like any Bollywood star, I’ve changed my appearance as you can see. That’s part of how Bollywood works. You have to have multiple costume changes. So here’s mine. I’m getting used to this air meets technology, it’s terrific and I love how this has come together so far. Thank you for watching for the last 35 minutes or so, and we have a few minutes for questions. So thank you for your comments along the way. I try to keep up, I don’t know if I was able to. And we have some construction work going on in the house, so if you hear some noise, I hope it won’t be too distracting.

I’m going to read out the questions that are in the Q&A. If you put a question in the chat, please make sure you put it in the Q&A and upvote the questions you want me to address for sure. So again, thank you for your attention. I said the scarcest resource is human attention and you gave me your attention. I’m very grateful for that. Here’s a question. If you’re a small institution, what are your recommendations for allocating resources, time, no money, et cetera, to understanding this as a tool and experimenting with it? This is a question that has come from the chat. These may or may not be by Eileen who’s on the staff. So this is a great question and something that we should all be thinking about. I can tell you having worked at one of the world’s largest museums and then worked with other large museums, that no one ever has enough budget.

No matter what your endowment, how many employees, they’re always short-staffed, always less than what you would want. And therefore, I think the best way to deal with understanding this technology is experimenting, is doing pilots, is trying things here and there without using it for mission-critical work. And there’s so much good stuff happening out there. So many people in the museum world that are experimenting and then sharing out what they’ve learned that we can all benefit from that. So I tell folks to keep an eye out for the conversations that are happening, including those led by the AAM and seeing where it makes sense for you to try stuff. It might be in the marketing side, it might be in dealing with collections, it might be in trying to understand revenue.

All of these are possible. The endless emails that you have to write in dealing with trustees, with donors, with members. So start small and try to find the things that make sense for you. You don’t have to do everything. And this is still very early. That’s one of my messages is that we’re very early in this process because it’s only a couple of years old. And so whatever your size of museum, experiment with it, and you’ll land where you need to. So then you can decide how much resources you can give to this at some point if it makes sense. All right, let’s keep going. Question from Kimberly. “Is there anything that should never be done by AI in museums? What should be codified in policy?” What a great question, Kimberly. Thank you. And I think the way that I think about these kinds of questions is that it’s hard to say what should never be there.

I want to go back to the point I made that this should be about humans, about the soul of what we do. And every day if you have any kind of public presence on LinkedIn, you’re getting messages, people trying to sell you stuff. And a lot of those, the volume of that has intensified because so much crap is now coming through from the folks who are creating all this content using AI. That’s one of the things that as it has made our lives a little easier, it’s also made the lives of spammers and scammers and others easier. So the guardrails should be that the things that your audience will expect to have been touched by a human, surfaced by human, shared by a human should always be without AI as the main thing you’re doing. It can be on the backside, it can be on the foundational level, it can be something that gets you started, but everything needs to be double checked, triple checked.

And the level of forgiveness that people have for seeing AI stuff, the tolerance will rise. But it is so early that there will be people who will not forgive you for using AI and pretending. So it’s all about how authentic you can be. And if you use AI, admitting that you use AI I think will make a difference as well. I’m going to keep an eye on the chat here so that Mike can let me know how much time we have left. Let’s keep going to another question. Let’s see. Question from summer. “As trusted institutions, what is our responsibility to reveal AI use? Obviously a generated image or final text would need to be attributed. What if AI is used for a first draft, a title?” That’s just exactly what we were talking about. Thank you. Summer.

I think what we would want to do is to make sure we can reveal that when it’s, I would say a preponderance of the work is done by AI. If it’s just a thought starter, if it’s just getting us out of the block, things like that, I wouldn’t worry about it too much. But it really depends on if you were to say, “This is really my work versus this is mostly the work of the computer.” Then you need to say it. But what we don’t realize is how much of this is already coming at us via this content stuff that’s coming. And so that’s how I would think about it. I’ve been given a little over a minute, so let’s keep going. Question from Linda. “How should museums balance AI use knowing that there are climate impacts that disproportionately impact people?” Thank you. And that’s that point that we need a trillion dollars or more fresh, new renewable energy.

Where the heck are we going to get it? So understanding that you have to decide that using AI is going to actually help you and is going to help your work get smarter, get better, more efficient. You’ll be able to serve more people, then that’s the balance that I would look at, if that makes sense. There is no clear answer on this because even Google searches use AI now. So every time you’re using a Google search, there’s a cost to it. So we have to find that balance and see what else we’re doing. I remember once talking about impact of museums and renewable energy and all of that, and somebody said, “You leave the lights on all the time and there’s so many lights.” So there are a lot of lights on. So what are other things we can do to reduce our carbon footprint, I think is going to be important. From Kara is a question, “Really enjoying the talk.” Thank you.

“Thinking about AI as a badly behaved teenager, scraping whatever it likes from the internet, how do we best discipline it to make sure that cultural data information from museum websites is safe?” Thank you. Great question. The way to do that is to make sure that you don’t start putting things on to ChatGPT and all of those places automatically. Think about creating a sandbox where you can have content and you can experiment with things without it going completely onto the internet. So that’s something that you should definitely play with. And as you think about policies, strategy, scenario planning, those are things that I think a lot about and I’m happy to help you think through what the questions should be for that. We have just a little bit longer. Let me see a question from Heather.

“What role should museums play in helping artists or writers protect their creative product from AI exploitation? Are there tools that are not ethically compatible with the core mission of museums?” It’s a great question. I don’t know the answer to that. I have not looked for that answer, but I’m going to work on it. You have in the chat, I posted my Gmail address. Please reach out to me. Otherwise, find me on LinkedIn, Instagram. I’m happy to work with you on that. You mentioned that designers need to use how to use AI. Can you expand on… Designers, yes. So I think design people are seeing how much AI is in the creative space and putting things that may make it harder for us to kind of have clear guardrails and design guidelines that we need to share.

So I would work on ways in which you can maximize the use of your intrinsic AI skills, right? Now, Canva offers something with AI, but there’s nothing that an AI design expert… Sorry, a design expert, they cannot top what an AI design expert can do. Not AI, no, I keep saying. A design expert can do. So one of the things I’ve always said about technology, I would say when Photoshop came around for the first time or when InDesign or tools like that, Adobe, QuarkXPress, these are all words that some of you may not even know. We would say that you can teach the tech, you can teach the toys, you can teach the platform, but you can’t teach taste, you can’t teach style. And that’s where designers and others who are really creative and know this stuff so well can have a huge impact on what they’re doing.

I also have a group where we talk about AI, a WhatsApp group that’s also in the chat. I hope that helps. I’m out of time. Thank you very much everybody. I’m Sree Sreenivasan. Super grateful to be with all of you. Thank you. Thank you. Thank you. Bye-bye.