Human First, AI Forward: Navigating the Future of Thought Leadership | Jen Cohen, Stephanie Grayson, Amelia Ross, Spencer Ante | 656

Welcome to Leveraging Thought Leadership. I'm your host, Bill Sherman. In today's special compilation episode, we dive into one of the most transformative forces of our time, artificial intelligence, specifically how it intersects with thought leadership, business strategy, and innovation. You'll hear from 4 expert voices we've recorded over the past few years. Jen Cohen, Stephanie Grayson, Amelia Ross, and Spencer Ante. Each of them will be identified by the role they served when we interviewed them, and each brings a unique perspective on how AI and large language models are being used, what benefits they offer, and where caution is still needed. We begin with Jen Cohen, who is VP of Core Engineering at the Toyota Research Institute when we spoke. Jen believes the future of AI isn't about replacing humans, It's about amplifying us. You mentioned something that I want to explore a little bit deeper. You were talking about the concept of human amplification and helping us do more through automation. And you've mentioned a term, ikigai. What is ikigai there? And how does that fit into some of the work the institute is doing? Because I think it's an interesting concept. On its own from a thought leadership perspective?

I love the concept. So studies of ikigai teach us that we feel most fulfilled when our lives incorporate work that we love and help society. So one of the things we think about is how do we enable people to achieve their ikigai? And a lot of that we're thinking about in relation to automation with a human touch and human amplification we talked about. Let me give you an example. I think it was 2 years ago, we had a slalom track set up. So, you know, track with lots of turns and curves and done with cones. And they had a lot of us drivers get in the car and they said, go through the cones and go up to 30 miles an hour and try not to hit any. I didn't get through. I got 25 miles an hour and I still hit cones.

Cones.

Yeah. Yeah. Then they turned on the automated driving and it was an assist. It didn't drive the whole car. I was still driving, but I could go 30 miles an hour all the way down to the ground and it helped me turn. It assisted me. It amplified my skills and I didn't hit one cone. Cone. And one of the things, um, our CEO talks about is the joy of driving. I've never had so much fun because I had no risk. I wasn't going to hurt anybody or the car, but I got to go fast through those cones and not hit a single one. And I think that's human amplification, not human replacement.

And that leads to the question of if you take out the risk or if you take out the danger, and whether that's, you know, aging in place or driving in that, what new opportunities open up when we reduce the risk of negative consequences? And you talked about joy, but I think there's a number of human emotions. And you talked about the empathy, you know, on the work and the projects. There's a very human part tied not only to the research that Toyota Research Institute is doing, but also the problems that are trying to solve, right?

Absolutely. When I think about the elders that are aging in place and looking at how much care our researchers take in interviewing people and understanding problems and understanding that ikigai is important in this, that people don't want just to have a robot running around the house doing everything, people wanna be involved. And so this idea of having the robot help to make the meal, having the robot, but, but still the human being able to be a participant in that, I think it's such a, a key part of how we can leverage that empathy, that we can understand that people need purpose and it's not just about replacing. And so yeah, I definitely think that that's in line with the way that we think. And if it's automated driving, you know, every bit of safety we have out there doesn't just help the person in that car, it helps all the people on the road with that car. So I love that that's, that altruistic part is key to TRI. That's one of the reasons I work there.

People who I see working on thought leadership and research development, that touchstone of why they do it becomes critically important because it's not just knowledge for knowledge's sake. It's about solving problems that people encounter that have a material impact in the world. Next, we turn to Stephanie Grayson. She's an experienced B2B content marketer and a seasoned thought leadership practitioner. She brings a unique perspective on how AI intersects with content and thought leadership and where its strengths and limits lie.

One of the most important things about thought leadership is that it gives you pause, right? Great thought leadership should stop, you know, really stop you in your tracks for a moment and say, now that, that is an interesting idea, or that is an interesting thought. And so just circling back to the question you asked about, you know, sort of where AI will be helpful. And of course, you know, to your point, I'm speaking about AI where we are today in 2023. I do not have a crystal ball, although, you know, it is interesting because I too am a consumer of thought leadership, reading, uh, up constantly on where people, you know, very deep into data science and the AI space are thinking where we might go and soon with the further development and sophistication of AI. But speaking as we are today in 2023, I think that it is for this reason, uh, that, you know, thought leadership is this conversation starter where we're thinking of something novel. It's not just a reorganization and repetition of things that are already out there, right? There's nothing I can kind of feed into, like there's no, there's no training data that the generative AI will have access to if this is a really fresh idea until you feed it that idea. Right. Right. But these, these ideas that are coming out are originating with humans. And yes, of course, if you have great ideas and you give them as training data and other types of things for generative AI, it can learn and it can learn to reason, uh, and, and things like that in the, in the future. And even now, right, we're starting to see elements of that. We're starting to see that even though ChatGPT is the media darling, right? We're starting to see also a lot of excitement around something called AI agents, right? So you have things like AutoGPT and other types of things where instead of being task-focused, it is, it is more of a suite of things where it can prompt itself, right? So let's say ChatGPT, you might have a human who's prompting AI, hey, can you, you know, suggest a listicle of 5 things about a certain topic or et cetera, et cetera. Whereas if you're working with an AI agent, instead of you prompting it for every little thing, you're going to kind of give it, let's say, an, uh, a goal. And then the goal means that the agent will then break it down and prompt itself into, you know, but still has to go through the process. And you see it kind of, you know, quote unquote thinking, right? About each piece, assigning itself each prompt, and then executing on each prompt. And I think in terms of the future, that's gonna be the interesting thing that we'll want to watch. And also that is a concern as well because things like bias can creep in and will, right? Because if it's learning from humans that, you know, we are not perfect, right? Humans are not perfect and come with their own set of, uh, you know, bias. Also, you know, that AI hallucinations, as they're called, is, is an issue where And AI can very confidently, very confidently, in fact, I mean, if you read it, sounds like it's sure, hallucinate things that either haven't happened, right? You have that lawyer who tried to use it where it cites—

That's exactly the, the example I was thinking of. Yeah. So go ahead and, and tell that story for our listeners.

Don't exist, right? A, a lawyer tried, lawyer erroneously tried to defend a client by citing cases offered that to him regarding case precedents that really didn't exist, but the AI generated them very confidently. That's something we need to—

And then the judge said, hey, your opposing counsel can't find these cases. Can you provide the transcripts? They went back to the AI. The AI generated the transcript or the documents. And made them look real. So they went 2 layers deep down a rabbit hole of hallucination, right? And it wound up being in some real-world trouble of having to appear before the judge and explain why they thought this was a good idea, right?

Yeah. And I think it, it really speaks to why human involvement will continue to be now and in the future, a very important thing that, you know, the humans working with AI, right? Augmenting with AI. But not replacing themselves with AI completely because who is minding the store here, right? I mean, I know personally when I've used several different AI tools, generative tools that I have had, for example, links that where I go to click them and I'm getting error pages or that they don't exist. So you absolutely must have a human in the loop because thought leadership done well can mean amazing things for business. Thought leadership done poorly, or if there's inaccuracies, you know, can really ruin reputation in a big way if people feel that they can't trust what's coming from the company in terms of their thought leadership.

Well, and from a thought leadership creation perspective, I think one of the things that you can use AI now, at least some of the tools is as a research assistant. If I were to put it in an academic setting, I would say, go give me an annotated bibliography of what people are saying out there and who are the voices and what are their, their points of view and take. I think you can task it with that to get a good snapshot of an ecosystem of conversation, but you then have to validate and go, is this real or hallucination? Who do I agree with, where are they sort of going, what is their direction? But there are limitations. It's sort of like having an intern or a research assistant at this point.

You know, look, even if you weren't a research assistant, there are reasons why we need one, you know, not just one person minding the data and minding the resource. You know, whenever you have an important piece of thought leadership, that is being put out there by a company, I would hope that more than one person looked at that. And I—

Absolutely.

I would hope that more than one person checked on it. So I think in general, AI or human, really, if you're working on an important piece of thought leadership, a few different sets of eyes should be looking at that.

If you're enjoying this episode of Leveraging Thought Leadership, please make sure to subscribe. If you'd like to help spread the word about our podcast, please leave a 5-star review at ratethispodcast.com/LTL and share it with your friends. We're available on Apple Podcasts and on all major listening apps, as well as thoughtleadershipleverage.com/podcast.

Next, we talk with Amelia Ross.

Director of Content Strategy at Primer AI. Amelia specializes in making complex AI concepts understandable and actionable. You talk about a little bit of being controversial and creating a conversation around something that may not be on people's radar. And so one of the things that I know with your organization recently has captured attention as we're recording here in May of 2022 with the war in Ukraine. Can you use that a little bit as an example of how that attracted attention and got people thinking about AI in a new way?

Right. My company Primer AI does artificial intelligence and natural language processing. And what natural language processing is, is essentially taking unstructured data and being able to run AI models against it. And so what is unstructured data? Unstructured data can be anything that's written, anything that's a video, et cetera. So you could imagine that in really rapidly evolving events like the Russian invasion of Ukraine, So much information is coming in, in non-written form. So what we did in our terms of our thought leadership is we took some of our products, one product in particular called Command, and we basically fed it information about what was going on in real time in the Ukraine. And we then created thought leadership content around that. So why is real-time situational awareness. So key for not only government organizations who are involved, but for corporations who might have employees or facilities in areas of risk. It turned out to be really a good way for us to engage a much broader audience beyond who I would say our initial cohort is, and to have In fact, reporters, many reporters coming in and asking to, you know, find out information through our services about what was really happening all too quickly.

And that ability to process real-time but unwritten information and to make sense of it. I think what you talk about as you're reaching different audiences, in some ways you gave a concrete example. Of something that is now possible that many people wouldn't have thought possible. Is that correct?

Yes. I would think people who were in government agencies would know this is possible. And I think people in the sort of on the cutting edge of natural language processing would know this is possible. But many people, sort of everyday citizens, would not know the extent to which this is possible and to the extent to which this is possible in a way that creates safety for citizens and safety for employees. So I think this was a very good theme for us to communicate broadly. It helps communicate the possibilities of NLP or natural language processing in a way that is maybe less fearful about, you know, and You know, it's also very engaging because one of the things that I think we all face as consumers of information is the amount of disinformation that's in the world. And we really use some of our technologies to surface what, what could possibly be disinformation. So I think, you know, we touched upon a number of themes that many people in the world are considering when they take a look at artificial intelligence and with the state of the world today.

And with that, as you said, those in government, those working in AI natural language processing, they're more familiar and more comfortable with this. But if you're leading an organization, if you're running a business, saying, how will this impact me? And what we've been doing, let alone if you are a citizen in a society, how does information flow? That touches on some very broad themes that go beyond and comes back to that question of going beyond the economic buyer to influencing conversations within society.

Yes, yes. And I, I think you'll see Primer and you'll see other AI vendors talking about the democratization of artificial intelligence, where it really is something that can be pervasive in society for societal good.

And now to wrap up, we turn to Spencer Ante. Most recently, Spencer was the head of insights at Meta, and he has a background in business journalism and investigative reporting. He's been studying how generative AI is reshaping the business landscape. So we've talked about some of the past and some of the present and thought leadership. I want to push and go into the future. And you and I have had conversations around collaborative journalism and AI, right? And I think given what we've seen over the last 12 months, really on using AI to generate content. It's both fascinating. And then the question of where does it sit within the world of thought leadership? So let me ask you that question. What are you feeling? What are you seeing and how's it going to change?

Yeah, this is the, this is the big question that a lot of people are asking right now. And you mentioned the term collab journalism. That's a term like I came up with when I was at Edelman. And the whole idea was like, make the journalist your friend. Instead of your foe, right? Because a lot of companies have an adversarial relationship with the media, especially the bigger ones. And so it was like, make the journalist your friend instead of your foe by like hiring the journalist, getting them to tell your story. It's the same thing, I think, with AI. Make AI your friend, not your foe. Now, AI has been around for a long time. It seems to be reaching a new inflection point with ChatGPT and DALL-E. On all these like gen— so-called generative AI tools and services. And, you know, if you've used them, you could see that like, it's, they're able to do more than they've ever been able to do. Like I just—

they're far more than a chatbot, right?

Yeah.

I went in this morning and I showed my son. I'm like, what is the main theme of Animal Farm? Okay. I typed that in there. 30 seconds later, it starts spitting out this like very like expertly composed paragraph. Okay. That was pretty good. But listen, the whole, the whole fear of AI is that it's going to replace humans. Now, I do think there will be certain situations where AI does actually take over like some lower level jobs of individuals that, um, in particular that where there's like rote or very like commoditized work. But on the whole, it's more an issue of society figuring out how to collaborate and and literally supervise AI, that's going to be the next 10 to 20 years, right? Because if you've ever studied AI, you know that all AI needs to be trained. Okay. Well, who does the training? Humans lead the training. Humans tell the AI what to, uh, what to learn, what to ingest, what information to absorb. And so if you, you know, it's the old computer saying, garbage in, garbage out. If you put garbage into, into ChatGPT or other AI, it's going to produce garbage. And, you know, there's a lot of instances of where you put in questions to these, these tools and they come back with like essays that have really key factual errors and inaccuracies. Okay.

And that's exactly something that I wanted to talk to you about because you mentioned the concept of being an advocate for truth earlier, right? And if your AI and the algorithm is working from information, which is not accurate, which is not valid. It can argue incredibly persuasively for something which is flat out wrong, right? That it, it can create the best essay in the world or the best campaign, but start from flawed premises that every human would look at and go, that's not right.

Yeah. And that's exactly why I'm not scared of AI, to be honest with you, for the most part, because I know having worked with it and studied it, that it's going to require human oversight and supervision for a while. Now, who knows 30, 40, 50 years down the road, these systems do get smarter. That, and that is like one of the brilliant things about AI is these neural networks and machine learning systems. They do get smarter over time, but that could take a long time. Right. You know, there was one example, I think somebody wrote an article that like you put it like who created the iPod and it says like. John Sculley created the iPod at Apple. It's like, anyone who knows Apple knows that's not true. It came— Jobs was there. So like, when I was at Bloomberg, one of the dirty little secrets of working at Bloomberg, the reason they were so fast is because 60 to 70% of the articles was regurgitated information from older articles. They basically like copied and pasted like what the, you know, the, like, you know, in an earnings story, you would take a bunch of stuff from another story.

Well, you knew what the template was going to be on an earnings story to begin with.

Yeah.

That's exactly right. That's like a fan of automation. So like, this has been going on for a while. It's just kind of reaching like a new level of sophistication, but still human beings need to like, need to manage it. And so I think that a lot of jobs are going to involve, um, in like, if you look at the job description of something going forward, it's going to say working with AI systems to like manage them, blah, blah, blah. And you're going to see a lot more of that.

It's going to be more similar to, you know, do you have basic word processing and spreadsheet skills, right? Can you manage those? Can you formulate a query? Can you create a dataset that's going to allow you to answer the questions that you want to answer? Right. And so the ability to curate knowledge, I think even if we go further back, you know, I think of my father who spent most of his career as a professional librarian. And he said, you don't have to know all the answers. You just know how— need to know how to to find the right answers and which sources to trust. And I think that that's the same response we could use for AI, right? What do we feed the system so that it has valid information? And then two, how do we evaluate it?

Yeah.

And you know, there are obviously technical advantages to these systems that human beings can't replicate, right? Like healthcare's a really interesting example. I think, I think, I think AI is gonna play an increasingly important role in healthcare because If you're a doctor, you need to stay on top of all the research in your field. And that's not easy to do when you're serving patients all the time. So like, and this is, this was the promise of Watson, which never really came to fruition. But if you could have like a medical assistant that could basically quickly tell you what the state of the art of the research is on this particular issue, you could augment human intelligence. And I think that's what AI is going to do. It's going to augment human intelligence, but not replace it because you still need— the idea of empathy is a really interesting thing. Like machines don't have empathy really. Do they have creativity? They're getting more creative because they're just absorbing all this information and coming up with more creative responses. But I think the realm of true creativity and original ideas will still be in the realm of, of humanity.

Whether you're designing robots, crafting content strategies, or building enterprise software, one thing is clear: AI is changing the landscape of business, leadership, and thought leadership. But with change comes responsibility. And as our guests have shown, thought leadership isn't just about evangelizing technology, it's about guiding it with empathy, strategy, and ethics. Okay, you've made it to the end of the episode, and that means you're probably someone deeply interested in thought leadership. Want to learn even more? Here are 3 recommendations. First, check out the back catalog of our podcast episodes. There are a lot of great conversations with people at the top of their game in thought leadership, as well as just starting out. Second, subscribe to our newsletter that talks about the business of thought leadership And finally, feel free to reach out to me. My day job is helping people with big insights take them to scale through the practice of thought leadership. Maybe you're looking for strategy, or maybe you want to polish up your ideas, or even create new products and offerings. I'd love to chat with you. Thanks for listening.