Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

The Media Copilot

The Media Copilot artwork

Inside Time’s AI push: Mark Howard on building an agent, not just another widget

Nov 21, 202551 minAudio
ShareXLinkedInFacebookWhatsAppEmail

If you grew up with Time magazines lying around the house, you probably have a sense of memory of that red border: serious, premium, grown-up news features - not just the facts. Now imagine that same publication trying to stay trusted in a world where people increasingly turn to AI for that kind of summarized analysis of what’s happening in the world. In this episode of The Media Copilot podcast, Pete talks with Mark Howard, Chief Operating Officer at Time, about what it actually looks like when a 102-year-old publisher decides AI is not just a threat to manage, but a system it needs to understand from the inside out. Howard describes a shift that feels less like a tech pivot and more like a newsroom learning to breathe in a new atmosphere. Instead of blocking AI or pretending the moment will fade, Time chose to engage, studying how bots interact with their stories and eventually creating its own AI experience for readers - grounded entirely in a century of reporting. The interesting part is not the shiny features. It’s the posture. Howard talks about AI the way editors once talked about search or social, an environment you may not prefer but one you have to understand if you expect your journalism to be found. The outcome of that work is the Time AI Agent, built with Scale AI, which pulls only from Time’s archive rather than the open web. It is a structured doorway into Time’s entire archive, built for an era when readers want context delivered instantly, in their language, in their format, without losing the integrity of the reporting beneath it. Howard notes that the core act of journalism is still carried out by reporters, editors, and the relationships they build. The work that appears in Time comes from interviews, access, verification, and context, not automated output. The Agent is meant to help readers move through that reporting in different ways, not to reshape or replace the work itself. What is most compelling is why they are doing it. The belief that the next phase of the internet will center on agents interacting with agents. And if that is where information goes, Time wants to make sure its archive can be read in that new language. What we cover in this episode • How Time decided to negotiate with AI companies instead of suing or sitting on the sidelines • The behind-the-scenes plumbing needed to protect IP and understand bot traffic • The evolution from Person of the Year experiments to daily AI audio briefings to the full Time AI Agent • Why Time built a private model grounded only in its own archive and what that means for accuracy and trust • How Time is thinking about AI marketplaces, enterprise licensing, and the emerging agent to agent web • What all of this means for the newsroom, reporters, and the future of audience relationships If you want to go deeper Mark’s articles on Time: https://time.com/author/mark-howardThe Story Behind the TIME AI Agent: https://time.com/7332572/the-story-behind-the-time-ai-agent LinkedIn: https://www.linkedin.com/in/markdhoward X: https://x.com/markdhoward This post was drafted with AI and then carefully edited by Media Copilot editors. 📩 Enjoyed this episode? Subscribe to The Media Copilot on Substack, Apple Podcasts, Spotify, or your favorite app. On YouTube? Tap the Like button and Subscribe to the channel. For more AI tools and resources built for media professionals, visit MediaCopilot.ai. Produced by Pete Pachal and Executive Producer Michele MussoEdited by the Musso Media Team Music: “Favorite” by Alexander Nakarada, licensed under CC BY 4.0 © 2025 Musso Media. All rights reserved. © AnyWho Media 2025 This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit mediacopilot.substack.com/subscribe

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

SPEAKER_02We have 102 years of telling the story of the history of the world. And a lot of the ways in which I'm hoping that people will want to use it is to contextualize what's happening in the world today to other historical figures or time periods and being able to draw and build that connection and have our agent create that output for them to be able to compare.

SPEAKER_01Hello, and welcome to the Media Co-Pilot. It's a podcast that explores how AI is changing media, journalism, and content creation. I'm Pete Paschel, longtime tech journalist and founder of the Media Co-Pilot, and I'm excited to bring you fascinating conversations with fellow journalists, media executives, and the people building how the business of news will work in the future. Today, I'm joined by Mark Howard, COO of Time. I was really excited to talk to Mark, partly because I've been reading Time most of my life. I remember clipping movie reviews from the magazine back when I

was a kid,

SPEAKER_01and that distinctive red border, which Mark and I touch on, has been a really effective symbol for Times brand of premium journalism. But now, with AI in the picture, there's a strong sense that premium journalism is under threat. AI portals have a tendency to flatten information, summarizing it and turning it into something that feels more vanilla and brandless. Nonetheless, all media brands need to contend with AI since it has clear utility, and consumers are using it more and more as an information tool.

SPEAKER_01To help ensure Time has a voice in how its content is used by AI, the company has been very aggressive about introducing its own AI experiences. This started about a year ago around its annual Person of the Year package, which introduced new tools for summarization, translation, and exploring context. And now Time's taking its homegrown AI to another level, what it's calling the Time AI Agent. Yes, it looks like another chat window, but this one can access the entirety of Times Archive,

SPEAKER_01all 102 years of it. And the information's grounded, meaning it isn't junked up with the entire internet, just access to Times content, all of it, of course, human-written and human-vetted. Times betting that that has some value, and not just to the people visiting the site. By putting an AI layer over its archive, time has essentially turned it into an AI product, one that can be easily slotted into someone else's systems, for the right price, of course. Mark and I talk about this potential new business model,

SPEAKER_01the factors that underpin time's big AI push, and where he thinks it's all going as more and more internet activity becomes agents talking to other agents. Hope you enjoyed my conversation with Mark. And if you do, I'd be grateful if you'd leave a review for this podcast. Reviews, especially in Apple and Spotify, really do help the show. So please leave one of those or a star rating if you can. And a quick note, you might not know this, but a big part of my business at the Media Co-Pilot is teaching AI to editorial teams. I host a monthly virtual AI class for beginners, too.

SPEAKER_01It's only 60 minutes, and like everything I do, it's focused on practical takeaways for media professionals, people in journalism, PR, communications. Plus, it's only $49. Each one's a live class, so it's always up to date, and I guarantee you'll walk away with something you can apply to your day-to-day. You can sign up at mediacopilot.ai. Okay, now it is time for my conversation with Mark Howard, COO of Time, Inc. Mark, welcome to the Media Co-Pilot.

SPEAKER_02Hey, thank you so much for having me.

SPEAKER_01No, it's our pleasure. So I'm really excited to dive into all these things with you. Just a little quick personal note. It's kind of a thrill for me. My parents were Time subscribers when I was growing up, and there were Time magazines all over my house. So I always sort of had this association with Time as kind of like premium journalism, right? That's wonderful.

SPEAKER_02Actually, you touched on something that we talk a lot about. We get the great privilege of traveling the world for time. One of the amazing things that we discover is so many people along the way have their own personal time story. And it's everything from people learning how to English by reading it as a child, to learning how to tell stories and having a broader perspective of the world, to even had a driver at a conference that I was at in Barcelona

last year who told

SPEAKER_02us that as a kid growing up in Nepal, they used to have Time magazine delivered to the local library. He would go there every day or every week when it arrived and read it. And it was his entry point into the rest of the world. So we love that. We hear that people have these time stories. And we think that with that comes a tremendous amount of responsibility, being the operators of this tremendous brand at this stage of its of its lifespan.

SPEAKER_01Yeah, and that kind of legacy brand is obviously rare. I could get into the whole brand of it, and I honestly think the red hue of it is actually a big factor there, and that it's such a strong color that humans sort of associate with that I think it really resonates. But this isn't the branding podcast. This is the Media Co-Pilot. We want to talk about how AI is changing media strategy, and I think... For such a legacy brand and particularly one that is always sort of retained at premium quality,

it's very challenging time with AI in the picture because it has this sort of

SPEAKER_01flattening effect on information. Right. So I'd love to start there with you and sort of like AI is kind of this force and maybe you could just take me through in your time at the publication. Give me the really quick, what has the conversation been around AI and how you have developed what the times approach is? And obviously, we're going to get into everything with the agent and all your recent stuff. But I really would love to get a sense of how the conversation internally has evolved.

SPEAKER_02Yeah, this goes back essentially two and a half, I guess almost three years ago now to when ChatGPT launched and forever changed the world and certainly changed the conversations of what it meant to be a media owner, media creator, owner of vast libraries of IP, but certainly more importantly, the arbiter of information to consumers and the different delivery mechanisms that we had been been living in the reality of the Google-based open web that we

had had for nearly 20 years.

SPEAKER_02At that point, it became pretty clear that there were lots of challenges and also some opportunities for media companies to be thinking about. where we were at. And Jeff Sibley, our CEO, decided a long time ago that there were three paths forward. You could litigate, you could negotiate and take more of a partnership style approach, or you can do nothing and wait and wait and see and see how things materialize, see how opportunities materialized at that point.

SPEAKER_02And she made the decision that we were going to negotiate. And that was for a number of different reasons. First and foremost, we understood that we as an independent standalone media company were not going to be able to navigate going forward on our own. And so it became critically important for us to take this partnership based mentality of If we can become a partner, if we can have a seat at the table, if we can leverage access to the brilliant executives and the

tremendous product

SPEAKER_02people and engineers at so many of these different technology companies, that we would gain insights and have information that we would otherwise never be able to accumulate just by being very active observers and students of the business. That opened up a lot of paths for us. And what ultimately ended up happening was we classified things into really three different buckets. There's the infrastructure layer, which is truly understanding in terms of how to protect our IP.

SPEAKER_02One of the first deals we did was verify with Fox. We were their first partner that plugged in at the CMS level and enabled us to then put a documentation of every single piece of content that we were generating onto the blockchain. And this is, safeguarding us. This is future proofing us for down the road. Just having that ledger and knowing that we've got that data protection is a first important piece. On top of that, we quickly followed suit by partnering with both Tolbit and

then Scalepost to do

SPEAKER_02more infrastructure layer tracking of all the bot activity on the site. And I've got a ton of interesting, I think, information around how we've seen that change over the last two years. But just laying that foundation so that we could track every day, every week, monitor what was happening, who was coming to the site, for what purposes, what articles they were going to, what articles they were coming back to, whether they were going for training purposes, whether they were going for indexing, whether they were going for AI search became a really important

part of our

SPEAKER_02operations just from an understanding perspective. Then, of course, came the licensing deals. And that's what's gotten a ton of attention in the media space. A lot of interest there. We were one of the early partners of doing a partnership deal with OpenAI. We also were early with ProRata, who I know Bill Gross has been on recently. So we're proud partners of theirs and continue to be admirers of the models that they're trying to build. We also worked with Perplexity and we worked with Amazon Alexa Plus.

SPEAKER_02So we've done a number of those types of deals. Those all come, they're all very complicated. They're all very confidential. They're all very different in terms of how those deals are structured, but each and every one of them most importantly to me, I think gave us access to not only the executives at each of those companies, but on a fairly frequent basis, access to the product folks and getting access to product roadmaps, to having engaged conversations with them about how they see

their products

SPEAKER_02changing consumer behavior, how they're leveraging that data to build their product roadmap out and some of the releases that are coming down and giving us access to I think really valuable insight that we're able to take back and have great conversations internally about the way in which they are shaping, um, the information flow and consumer expectations. So that's been highly critical. And then, and then the last part, which everybody is looking at is operational efficiencies.

Um, how can you bring these tools, different AI tools into your organization?

SPEAKER_02Um, There's a lot being covered, of course, around inside the newsroom. I don't run the newsroom, but happy to talk a little bit about some of our philosophy so far. We have not had some of the public announcements or pushed the newsroom in ways that others have, but we are certainly observers of how people are thinking about it. These inside the newsroom are certainly tools that people can use for varying degrees of helping them be able to do their reporting and to

generate their

SPEAKER_02content. But more so, I am responsible for how the rest of the organization is deploying tools. We can get into a little bit some of our approach to agents. around different job functions. We have some that are department, some AI, third-party AIs that are department oriented, and we have done a ton of work at the individual level to look at what are a lot of the repetitive tasks and responsibilities that people in our company have, and how can we either bring automation or create agents for

them to help them expedite a lot of that work so that they can do less of the manual work and do more

SPEAKER_02of the strategic thinking and the higher value work.

SPEAKER_01Nice. Wow. That was a very huge summary. I'm going to dive into a few of those bits here. So you talked a little bit about the multiple partnerships and agreements that you have with the various companies. I've had many of them on the podcast. And I'm curious a little bit about if you see any redundancy there. I mean, you must be seeing like a lot of similar data from various companies. partnerships, it seems like you're, it's almost like rather than choosing,

SPEAKER_01you're just like, let's get them all in here and let's find the best ones. Are there any that have been more revealing, I guess, about AI activity or even what AI experiences or even third-party platforms are most, I guess, attractive for you guys to learn more about or connect with more? Obviously, you're probably getting data from a bunch of different places. What's been the most promising?

SPEAKER_02Yeah, we are. So we've been both pretty aggressive in terms of meeting with a lot of the different companies that exist out there that are bringing solutions for media companies. And they take all kinds of different shapes and sizes. There are lots of them out there that I think are redundant to the 15 or so partners that we've already brought in officially into the fold. And we're continuing to look at more... One thing I didn't even touch on yet was the ads part of the business.

SPEAKER_02And there's tremendous innovation happening there. You just had firsthand on and they're a great example of some of the innovation. But there's a lot happening in terms of being able to deliver the business as well. So on the data side, yeah, there is a little bit of redundancy, but there's also, I think, a lot of unique and original data that's being brought to us by our various partners. But there's also a new path being forged forward. You just think about in the last couple of months, all of a sudden,

SPEAKER_02NL Web and MCP are important. And there's talk about AI marketplaces and there's RSL and there's other standards that are being discussed at an industry wide level. And there's a number of different companies out there that are really leading that charge. So where I sit as a media operator, we just can't possibly be doing all of that work on our own representing time, trying to forge a path forward. But these companies that are out there are fantastic at it and

they're specializing

SPEAKER_02in different aspects. of trying to help bring media companies like ours to the future, bring us opportunity. In number of cases, we operate as a design partner where they've got either proof of concepts that they're looking to develop or theses around what the market opportunity may be. We really like to lean into those. We are really great at the partnership level. We can talk on the agent side of how we've done that over the course of the last year.

SPEAKER_02But we think that there's a tremendous amount to learn right now. And we can learn by getting our hands dirty, by getting into the data, by having time data be part of the mix and really having these frequent, whether they're weekly, biweekly or monthly or quarterly meetings with our partners to dive deeper into everything. it's been a tremendous learning journey for us. And we've got a small task force of folks that are actively involved in that. And then we bring at the executive level every two weeks,

SPEAKER_02a higher level of debrief to what we've learned, what we're seeing as opportunity, Obviously, there's no shortage of threats out there that need to be discussed, but we're taking much more of a proactive approach and we view it much more as an opportunity at the moment that with the right partners in place, with the right approach to how we treat the time brand, both in the areas and domain on site where we can control, but also this notion of the loss of control when you go off

platform into these

SPEAKER_02other platforms where you don't own the distribution. That's certainly the truth, but the change of distribution mechanisms is nothing new to a 102-year-old media company like Time back in the days when it launched. 1923 you relied on newsstands and then over time the mechanism of information being distributed went to newsreels and it went to radio and then ultimately we had the invention of the internet and internet then as we all know well

chronicled went

SPEAKER_02from desktop to mobile to social and now to ai which most likely will be the most disruptive of any of those changes, but change is not new to time. And we've weathered that over the century now, and we are equipping ourselves to be best prepared to move forward.

SPEAKER_01Yeah, and you guys made the interesting decision to go forward with some of your own AI experiences. And I think that's a good time to transition to that because, you know, your analogies are very good and maybe not perfect, but like, you know, you talk about newsstands and obviously very key for the longest time for magazines for a long time. You didn't have too many magazines actually owning their own newsstands. Maybe they did here and there, but... So again, maybe this is early days of AI,

SPEAKER_01certainly, but the decision to do your own AI on site and first with, I think, officially with person of the year about a year ago, right? Then you have your audio experiences earlier this year and now with the agent. Before we go into the details of any of those, how was that decision made and sort of like, okay, we're actually going to offer our own uh take on ai um which not everyone has done some are sort of content to do take the data that you are talking about and sort of

adapt their content strategy or business strategy or tech strategy around it and just continue to do what they do but you guys are doing you know you're offering sort of brand new things so tell me about that decision and then let's let's talk about how those

SPEAKER_01experiences have evolved

SPEAKER_02Yeah, yeah. So Jess, our CEO, says all the time that we don't want AI to be something that happens to us. We want it to be something that happens for us. And with that, as sort of the backdrop, last summer, summer 2024, we decided to issue an RFP to 12 different companies, some in the traditional consulting world, some in the big tech world, some that were AI native companies. And we asked them, to help us imagine what an AI-based experience could be for the

time person of the year.

SPEAKER_02And of course, that was a pretty ambitious plan given the time person of the year is one of our biggest announcements of the year. Every year, it gets the most global pickup and attention and conversation. So to do it on such a high profile event was no small undertaking. We sent the proposals out. We got a tremendous amount of response with all kinds of really great ideas. But the one that really stood out for us at the end of the day came from Scale AI,

SPEAKER_02of course, a native AI company based in San Francisco. Pretty well known now, of course, given the meta relationship over the summer. But at the time, they weren't necessarily quite as well known. Alexander Wang had been on the Time AI 100 list and then later made the Time 100 next. So we were familiar with scale, but they came in with an idea that brought together a lot of the basic ideas AI functionalities of auto generating summaries and translation

and chat and putting it all together in a unified toolbar experience that we launched with

SPEAKER_02person of the year, not only last year with Donald Trump, who was the person of the year, but we went back three years and included Taylor Swift, Vladimir Zelensky and Elon Musk. And they created a very small model on all the content we had ever written about all four of those people. and launched the toolbar and really created a generative experience where people could go in and you can mix and match any of those functionalities. You could ask it to give you a medium length summary of the Taylor Swift person of

SPEAKER_02the year and have it read to you in Spanish. And it would do that on the fly. And then once that was done, you could then in Spanish or whatever language you chose, you could have a conversation with it chatbot style, but do it verbally. Or you could do it text-based. You could do any number of different things. And so this was us being able to take... In no way did it... changed the way that the content was written, that the article was edited. The original full-length article was presented on arrival for

everybody to consume

SPEAKER_02the way that they always have and probably expected to, except we greeted them also with this bar overlay to be able to make it much more interactive. And with that, we saw a tremendous uptake in the amount of time that people spent with the bar. We were pleasantly surprised that while that experience we had only introduced in the dropdown seven different languages, we saw over 20 languages were engaged with the bar. So people in their native tongue were just choosing to do it. And it gave us great confidence that we would be able to build

a experience on the

SPEAKER_02site that would be much more accessible to people in whichever formats and languages was best for them. And what I mean by that is I'm an audio learner. I'm an audio listener. I listen to your podcasts. I listen to many other podcasts. That is my information diet is coming in through podcasts. I have colleagues who are voracious readers and they want to read. So for me, I have gone to a model where in the past I would take

SPEAKER_02a long form article and I would put it in the 11 reader, the app by 11 labs, and I would have it read to me by Jerry Garcia. It's a voice that I enjoy and their, their app does a great thing. Great, great adaptation of it. That's my mechanism. So we saw with the toolbar at the person of the year that people were highly engaged. They were using the tool. They were either having long or medium length summaries generated from themselves and they were engaging with it. So it gave us the confidence to move forward with scale as our

partner to take on a

SPEAKER_02way bigger, more ambitious project, which was taking the entire hundred and two year archive of time. which sat in four different databases. Some of it just literally pictures that were converted to PDFs of magazines from the 20s and 30s all the way up to 20 years ago when we used the modern-day CMS to publish content to the web and then multiple iterations and varying degrees of complexity in between, they've taken the entire 100 into your archive.

They've indexed it so that it is now a private time LLM that is grounded only in our content.

SPEAKER_02There's no third-party information or content that is part of it. We've trained it up in terms of editorial voice, editorial guidelines, so that the writing is reflective of the way that time is conditioned to communicate with our audiences. And with that, it became the foundation of this basically year-long project this year that just culminated in the launch last week of the AI agent on the site.

SPEAKER_02Prior to getting there, though, we needed to have a quick stop along the way in terms of some of the audio developments. And so in June, one of the things that we did was release an AI audio brief that gave us the ability to really focus in on a core aspect of this experience, which is going to be the AI voices of time. We use Chad QBT as the underlying foundation model. I mentioned earlier, we have that relationship with OpenAI. So it became important,

but also a good opportunity to leverage that relationship and access to some of the

SPEAKER_02resources there to be able to build time product on top of. So in the summer, we launched a curated audio briefing with your hosts, Henry and Lucy, a nod to our co-founder, Henry Luce. And they then every day took the top four articles as determined by the newsroom and created this short form audio briefing explaining that day's news. That gave us the ability to fine tune the voices, the conversation, how they handle different topics,

SPEAKER_02how they handle transitions, and really tune it to get ready for the agent, which we have now deployed, which gives far greater functionality. Yeah, it seems to me, so the agent seems like the bigger step, obviously. Even though I will sort of throw a little bit of pushback skepticism to you. I mean, it's ultimately like kind of a chat window, a still chat window on an article, right? Which is like, you know, from an experience standpoint, it's very similar.

SPEAKER_01Now I'd have noticed in my use of the agent that, well, compared to what time AI was before, particularly with person of the year, Again, correct me if I'm wrong, it seemed like the summaries of those articles, while generated by AI, were also seemed like they were vetted. Like, in other words, whenever I would hit the generate button again and again on the same, whether it was Trump or Taylor Swift or whoever, I'd get the exact same text, like word for word, which suggests to me it wasn't being generated on the fly.

SPEAKER_01It was a vetted summary. That's different now. The Time AI agent now has, I've done it many times on several different articles and I get slightly different summaries each time. The facts are basically the same, but it's a different summary. this to me suggests you have a growing confidence in the technology to be able to do this, or at least with your implementation of it. Can you go into a little bit more about that? Because this is obviously gets at the heart of what people

worry about with AI.

SPEAKER_01And you get, you know, you could go back to the Apple summaries on BBC stuff that were saying completely wrong things. And this, this kind of grounding, I guess that you guys do might help prevent that. And, but yeah, Tell me about the confidence, how the technology has evolved, and what you still see perhaps as your duty in either disclosure or caveating to the reader.

SPEAKER_02Yeah, so you're exactly right that it is not cached summaries that it's generated at this point. It is truly generating a unique answer for each example. And one of the important things that we built into the agent, because the agent now sits on the article level, and we're currently in politics and entertainment, and we'll be releasing to other sessions on the site and eventually the homepage. The agent has the ability to understand the article that you're on.

And so when you expand the window and you get into asking it to summarize or

SPEAKER_02translate or find related articles, it does have recognition of the article from which you came from. In order for it to be truly agentic, it couldn't have cached information in there, nor did we want every single day with every article we published to write the one standard summary for each article. We wanted it to be dynamic. That's the whole point of it being agentic and not just an AI-generated summary.

So that was at the foundation of philosophically what we were trying to build and

SPEAKER_02what we were trying to enable with the agentic capability, but then also the ability, if you had a chance to look at some of the preset prompts that we also put on the page to be able to give people an indication of the more sophisticated actions that the agent could take on their behalf. Whether it was doing a roundup of all the recent coverage of President Trump and giving it to me in a five minute audio brief. Or we recently had an exclusive interview with the Prime

Minister of Italy, Maloney.

SPEAKER_02You want that given to you in a briefing in Italian. It can do that for you on the fly. So the ability to mix and match this functionality of being able to take the heart of the journalism in no way, manipulate the journalism in no way, influence the journalism, use the journalism at the core, put this agentic capability on top of it and enable accessibility to the information. However, people prefer to get their information in the form factor, in the lengths,

SPEAKER_02in the language. That became the enabling factor of what we were trying to do with this. Yeah, when I was trying to stress test it, I was really pushing to sort of older archive stuff.

SPEAKER_01And because I obviously have read Time for a long time, I could actually think about specific things. So I think my main stress test was sort of asking about... Actually, I forget what the query is now. But it was something about Tom Cruise and how influential certain movies in his... whole oeuvre were have been and you know asking about that because i remember like older features like i think richard corliss wrote them way back in the day in like the 80s or 90s and i was actually thinking about a very specific one and i went it wasn't gonna i was gonna name that one and it did it was like

the first thing that came up oh richard corliss wrote this cover story and i forget what it was 89 90 or

SPEAKER_01something and so so that got me thinking about that the can time be only one of the brands that does this because it has such a robust archive which uh is a bit i might be a bit flattering but i also see kind of the weakness in that too because it's also like well it is still just time right like as big as your archive is there's a big world out there and a lot of things that have been reported on so how are you thinking about this as both an advantage And maybe a disadvantage or maybe something to make up for with

other partnerships.

SPEAKER_01Tell me about how you're thinking about the person querying these things, the kinds of queries and the archive and what its inherent weaknesses and its strengths are.

SPEAKER_02Yeah, well, one of the things that Time has the permission to do, and we've always done, is we cover a pretty broad range of topics. And we've got authority and credibility in a lot of those different areas, which I think is great in terms of having this agentic capability to be able to draw from our archives. You're right that we do have an extensive archive. We have 102 years of telling the story of the history of the world. From a position of credibility,

SPEAKER_02of fact-checked information that we feel very confident in, it's rigorous editing over the years, even with the digital content that we published in the last 20 years, still edited, we still feel very confident in the information that is in the inputs that have made up the LLM here that we've trained on. We had almost three quarters of a million assets that went into the training of this model. So it's a lot in Internet standards that may not be considered

a lot by a lot of

SPEAKER_02media companies that may be quite small, actually, for a lot of media companies. But we do, again, feel that we've got this tremendous corpus of information of the history of the world and a lot of the ways in which I'm hoping that people will want to use it is to contextualize what's happening in the world today to other historical figures or time periods and being able to draw and build that connection and have our agents create that output for them to be

SPEAKER_02able to compare. I mean, how does... current administration terror policies compared to Reagan era policies. And you can easily have the agent do that for you and you can have it read to you or you can have a debate and discussion podcast style created for you that you can then get that information. So there's a tremendous amount that you can do. We are, as you mentioned, 100% grounded in time content. That was an important part of building this thing, launching this thing. I think like anything,

SPEAKER_02especially in AI, everything is about shipping products, getting it out there, getting the data, getting the feedback, seeing what people like, what they use, if they're using it, and then being able to continue to expand beyond it. A near infinite amount of ways in which you could see this product further developing. One of the things that is in development right now is we are going to be bringing sites registration back to time.com. Something that we did in early 2023 was we had removed the

paywall and we made time

SPEAKER_02open and accessible to everybody around the world. That was a big strategic decision. With the reintroduction of site registration, you can see where we could add memory to the agent, see where we could add much more personalization and customization. You could have delivery mechanisms of the information. We have some of the foremost reporting on the topic of AI with our our editorial newsroom, we have this franchise of the Time 100 AI list every year.

We've got this incredible wealth of information of what's going on in the world of AI. I could have a summary generated for me once the memories, once we add memory, generated for me that's delivered to me in my inbox every Friday or a weekly

SPEAKER_02podcast created to catch me up on all things AI. And those are just small examples of what memory could unlock. So we're thinking a lot about, is that the next stage of where you go? There's certainly, we think, one of the most valuable elements assets in the world would be all the covers of Time Magazine of 5,000 plus covers that we've generated over 100 years and being able to do some really cool things from a visualization standpoint. You could just see charts and graphs and more multimodal

experiences coming into

SPEAKER_02this depending on the types of questions and queries that people have. So there's a number of different ways with which we can go moving forward to greatly expand it In the near term, there is no plan to expand beyond time content, but you never know how things will evolve down the road.

SPEAKER_01Yeah, and I think a lot of the things you mentioned there, they speak to interesting experiences and potentially even like revenue sources as you go forward with user experiences. I was also thinking like in terms of what I think there's an important distinction here. And again, I think you should clarify this for me. But because you've developed your own essentially agentic layer on your own content and you've made it all AI ready, this is... sort of reversing the lens on a traditional licensing agreement, right?

SPEAKER_01Whereas like the traditional way, I guess, if you're a publicist, like, okay, go ahead and scrape my content for your AI engine. And that I would think is mostly just they're scraping what's online and it's probably not... quite as comprehensive and there's not a lot of control either because you're effectively just sort of handing over a lot of the IP. Whereas it seems with what you've built, you could just essentially get access to the agent in some kind

of MCP or sort of very simple sort of technical handshake kind of way because you've done all the

SPEAKER_01work. of digitizing and vectorizing and doing all this stuff, right? That's right. So what this opens up to me seems like, oh, okay, so now there's this archive of content that's historical, that's across a broad range of topics, that's all human generated and human vetted. uh going back over a century and it just seems like it would be i don't know something something pretty valuable to the right to the right customer who want wants easy access to that for whatever they're doing i don't

think they could be a bank or or a i don't know a movie like i'm just talking about tom cruise maybe some studios or something would want to do i don't know are are is this the b2b side of things is this something that a door has very consciously been open and are you

SPEAKER_01pursuing that

SPEAKER_02Yeah, everything is intentional in terms of being able to get ourselves ready for where the world is going, especially in the world of IP licensing and distribution. So like I mentioned, we have several partnerships already in place. We had to deliver or develop delivery mechanisms to be able to get them that content so that not only did they have access to the archives as part of those arrangements, but they had all the content that we publish on a go-forward basis.

SPEAKER_02This is exactly right. We've now vectorized all of it. We've indexed everything. We've created this agentic layer. We are thinking a lot about MCP, RSL, other standards and protocols that are being developed to be able to have us be more transactional on the agent-to-agent web. That's, I think, all in super early stages. We're also thinking a lot about AI marketplaces. We hope that a lot of the great work that Tobit and Scalepost

and others have done

SPEAKER_02to try to get the media side, the supply side, ready for AI marketplaces is While they haven't materialized at any great level yet, I know there's a number of other marketplaces that are well in development that potentially could be even larger on an enterprise scale. But the ability for AI bots to be able to quickly, easily, in a structured way, be able to find the information they want, retrieve it, hopefully ethically source that information,

SPEAKER_02and when appropriate, compensate fairly. A marketplace dynamic would be great, where obviously the grid prices are easy to transact on. That is where I believe this is all going. I believe that I know it's been well discussed. Bill Gross certainly talked a lot about this, but the ethical sourcing of IP and data to be able to be surfaced in AI answer platforms in any way. agentic experience. Your point is absolutely right. I think companies are going to be building all kinds of

different agents to go out

SPEAKER_02there and bring back information for their customers. So I don't think it's just the frontier labs. I do think that there's going to be a whole explosive world of agents out there that do need information on specific topics that they know and their ability to find it easily will hopefully be much more than worth the very nominal fee that it will cost them to be able to source that info.

SPEAKER_01Yeah. You don't want all the junk with it, right? And there's a verification layer to anything you're doing with AI. And the more you can have assurance in the quality of the answer and the quality of the AI getting it from that information, basically the more efficient that process is going to be.

SPEAKER_02Right. I mean, the big frontier labs have spent a tremendous amount of resources and money to build retrieval mechanisms to be able to go out there and retrieve the information, whether it's for training, indexing, or fetching for RAG purposes. Again, being able to help facilitate that and do so in a way in which it becomes much more efficient and effective for them to just simply go through one of these marketplace environments and to pay for it,

SPEAKER_02to source it ethically makes a ton of sense. And I believe that that is where we're all going. And to your point, a lot of this infrastructure level work is us getting prepared, not just for the consumer experience that we now launched on the site, which I believe and I hope will much more meet the moment in terms of the way that people's expectations of how to be able to get information exist as conditioned by the massive amount of people that are using ChatGPT every day, the massive amount of people that are using Gemini or

Perplexity or any of the,

SPEAKER_02or Claude or any of the platforms, right? That's now the expectation. I should be able to have that same ease of experience wherever I am. And while it's not necessarily native in a media website environment today, we're really excited about introducing it to people, I think, as they become more familiar familiarized with it, it becomes a little maybe less jarring when you arrive and there's this thing there that doesn't traditionally exist on a lot of media websites. But we hope we made it as intuitive as possible to be able to

get people in.

SPEAKER_02And once you're in there to show the functionality, so that people can start to sample it and get a sense for that experience. We're really optimistic that now we've done the front end consumer part, the infrastructure and the back end is in place. The agent to agent web is the next big frontier and being prepared for that and really having ourselves structured in a way in which we can be at the forefront of being able to operate is critically important to our overall AI strategy.

SPEAKER_01I know you mentioned you're not in charge of the newsroom, but I am curious about, you know, certainly all the stuff you're talking about and the new experiences you're releasing. Is there a certain level of, I don't know if it's skepticism or fear or, you know, from the editorial staff at all. I know that other places that have AI experiences and as they evolve them, there's been pushback from unions like at Politico, for example. What is the concern, if any, coming from the newsroom and how is your organization sort of

trying to adapt to

SPEAKER_01that?

SPEAKER_02Yeah. So again, I don't run the newsroom. And at this point, there's no mandates or directives to them in any way in which they need to be using AI to do their jobs. I think that my perspective is there's AI to do content creation, and then there's a broad swath of AI tools that exist today to help accelerate the pace to being able to publish, whether it's researching tools or copywriting or headline optimization or SEO

SPEAKER_02optimization. A lot of those tools have existed for a long time. And so, you know, a lot of it now is getting slugged as AI. I think those are just all writing assistants and people use them to varying degrees. Like I said, we're not mandating that the newsroom adopt any specific set of tools or workflows at this point. However, I do believe that most journalists at our organization and probably every organization are using various tools to various levels already and they're probably

SPEAKER_02only just going to continue to do so as they become more familiar with it. So there's no doubt that the way in which people are using all different tools to be able to generate content and to be able to slice that content for distribution to different endpoints. You yourself, you have a lot of different ways in which you can consume snippets of the information that's generated in this podcast and viewing in different platforms. And all of that is enabled by different tool sets. And so...

SPEAKER_02It's those tools, I think, that are going to become the main driver of change, first and foremost. The journalism, the journalists, at least for time, it's critically important we continue to support the individuals and their beats and being able to support them with the access that they have. We don't have reporters out there who are trusted. And that trust is their mechanism of getting access, whether it's the interview with world leaders or celebrities like Taylor Swift, who had not conducted an interview in over five years before

sitting down with time

SPEAKER_02for her person of the year interview. or the vast amounts of on-the-record time that our editorial team has with President Trump. That's all earned by trust. That's all earned by reporters who are top of their game, that have access to exclusive information and can tell stories that are unique to anything else that's being published out there. And that always will remain critical. That is what we're going to continue to do at the core of Time Capital J Journalism.

And that will sit at the center of any experience that we're able to enable by

SPEAKER_02putting AI either on top of it, like I said, for human-based experiences or taking that great journalism and preparing it for the agentic web.

SPEAKER_01Nice. Let's flip the lens again just outward really quickly. You obviously have a global audience. You're doing things in multiple languages. Are you seeing any differences in how international readers are responding to these features or maybe readers just using other languages? I'm curious what you're seeing so far and maybe what you're seeing in the very early days here in the ages.

SPEAKER_02Yeah, the initial data is suggested from the toolbar with person of the year that people were looking to engage with it in a number of different languages. What the agent that we just released will introduce is accessibility. We right now are available in 13 different languages. We'll expand that over time if we see that people are trying to engage with languages that it doesn't currently support. But just that accessibility component of anybody around the world being able to access any piece of content that we've generated and be able to

translate it. I know translation tools exist and people can do it if they're savvy at the browser

SPEAKER_02level or a number of other ways. But just bringing that onto the page experience and making that accessible right at that moment in time. I hope is a big unlock and we'll introduce as we continue to expand globally, want to be attractive to global audiences. As you mentioned in the beginning, the globally recognized iconic red border, it does travel globally. It is something that whenever we go around the world, we get recognized all the time. People have a

That speaks to the decades and decades of tremendous coverage that we've had,

SPEAKER_02the global reach that we've had. And this, I believe, is just going to be a further unlock for more people to be able to access it.

SPEAKER_01Well, let's, as we wrap up here, let's look forward a bit. And, you know, you mentioned the agent-to-agent web, and I'm wondering how you're thinking that's going to affect things, particularly with consumer agents. You know, you have this great on-page experience, but it seems that reader agents and personal assistants and, you know, the tools that people use, whether it's Notebook LM or something else, of just, you know, adapting content to themselves are going to be a huge... Taking all of that in, how do you see things evolving for publishers?

SPEAKER_01And I guess if you had to boil it down to advice you might give to other publishers, what they should be doing right now to get ready for this future, what would you say? Yeah, I can just speak to our own philosophy, which is we've really tried to learn by doing.

SPEAKER_02As I said earlier, most of the partnerships, the greatest value we get out of it is by having access and a seat at the table to learn directly from them. And then to take that information and to bring it back and to think through how we can best deploy for our own either internal operational capacity. growth or thinking about business growth through either how we communicate our content to audiences and ultimately the agentic web. The agentic web for sure represents a massive,

SPEAKER_02I don't even think we begin to understand the size of that market. It could be exponential and it may be infinite when you think about the deployment of agents on behalf of individuals and and organizations and companies that are set to be deployed over the course of the near term future. We believe that a very trusted source of information like time will play a critical role in how those agents can credibly come back and represent

the interests of the

SPEAKER_02user or the organization behind them. So that to me is, I think, a huge business opportunity. It's what we wanted to be prepared for. It's what we are preparing for. It's why we have as many partnership deals in places we do. And it's why we're going to continue to lean on them to help us navigate either as a thought partner, a data partner, or a design partner in solutions to be able to take advantage of that opportunity.

SPEAKER_01Nice. Well, Mark, this has been great. I feel like it's been part one of a 10-part interview, to be honest with you, considering all the sort of other rabbit holes we could go down with the terms of user experience, user agents, and all the other stuff going on in the space. But thanks for boiling down your thoughts on this and telling us what you guys have been up to. Absolutely. Really appreciate you taking the time and looking forward to parts two and three.

SPEAKER_02All right.

SPEAKER_01Take care. Thanks. Thanks for listening to my conversation with Mark Howard of Time. If you enjoyed this episode, I'd appreciate it if you'd take just a minute to leave a rating or review for the Media Co-Pilot in your podcast app. Also, don't forget to subscribe to the Media Copilot newsletter and check out mediacopilot.ai for our classes on AI for journalists, PR professionals, and communicators. Seeing you in the future.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Why Yahoo still matters and what it knows about the Future of News

    Jan 16, 202641 minAudio

    For years, Yahoo News has occupied a peculiar position in the media ecosystem. It is not a social network, not a traditional publisher, and not a search engine in the modern sense. Yet millions of people still use it as...

    Credit the people on this episode

  2. Best of the Year: Inside the AI shift that's transforming media and journalism

    Dec 28, 202531 minAudio

    This special episode of The Media Copilot is a look back at the conversations that defined the past year… the questions, tensions, and turning points shaping how media, journalism, and technology intersect right now....

    Credit the people on this episode

  3. Can you trust that clip? Storyful’s James Law on verification in the AI Era

    Dec 19, 202543 minAudio

    AI has made verification a newsroom survival skill. In this episode of The Media Copilot, Pete Pachal talks with James Law, Editor-in-Chief of Storyful, the newsroom partner known for verifying the videos and social...

    Credit the people on this episode

  4. Practical AI in the newsroom with Darla Cameron of The Texas Tribune

    Dec 12, 202541 minAudio

    AI in journalism can feel abstract until you talk to the people actually shipping products inside newsrooms. In this episode of The Media Copilot podcast, host Pete Pachal talks with Darla Cameron, Chief Product Officer...

    Credit the people on this episode

  5. Building the next era of the open web with Adam Singolda

    Dec 5, 202543 minAudio

    AI is rewriting the rules of digital media and few people have had a closer view of the shift than Adam Singolda. In this episode of The Media Copilot, host Pete Pachal has a candid conversation with the Taboola founder...

    Credit the people on this episode