Could AI ever replace the news?

Summary

Podcast Summary:

Business Daily (BBC World Service)
Episode: "Could AI Ever Replace the News?"
Date: September 17, 2025
Host: Sam Grouet

---

Overview

This episode of Business Daily explores the rapidly evolving role of artificial intelligence in the news industry. From AI-generated anchors and news summaries to the threat of misinformation and newsroom automation, the discussion probes whether AI could ever truly replace traditional news — and whether audiences would trust it if it did. The episode features expert guests including Adam Mossam (Channel One), Chris Stoker Walker (journalist and author), Dylan Jacks (Telegraph Media Group), and Akin Tunde Bapatunde (Center for Journalism Innovation Development).

---

Key Discussion Points & Insights

1. The Arrival of AI Anchors and News Content

• Introduction to AI-Generated Anchors:
  • In December 2023, an AI-produced news broadcast went viral, garnering over 5 million views in 24 hours ([01:22], [04:16]).
  • India Today introduced its AI anchor, Sana Mary, in the same year ([01:39]).
  • The show’s host, Sam Grouet, tries ChatGPT and Elevenlabs to generate and present a news script, demonstrating the speed and improving quality of current AI tools ([02:47]-[03:18]).
• Channel One’s Vision:
  • Channel One briefly launched an AI news service with content available in 30 languages, aiming for efficiency and regional relevance ([03:53]-[04:16]).
  • Adam Mossam: “I think we're operating on the leading edge of AI, and it was probably the first time a lot of people saw some type of digital avatar with that much precision or that lifelike.” ([04:49])

2. AI in the Newsroom: Help or Hindrance?

• Efficiency over Replacement:
  • News organizations face budgetary pressures, e.g., Business Insider laying off over 20% of its staff while moving toward more automated content ([05:03]).
  • Adam Mossam: “Our product essentially acts as an assistant to the creatives…helping them get a lot more done in the same amount of time.” ([05:21])
  • Channel One shifted focus to shorter, mobile-friendly content, revealing both technological possibility and generational habits ([05:59])
• Current Uses:
  • Chris Stoker Walker:
    • AI is already summarizing articles at leading news outlets, e.g., Bloomberg offers bullet-point AI summaries for readers ([07:27]).
    • “We are seeing AI being used on the front lines of journalism at the point at which the audience consumes that bit of information.” ([07:27])

3. Legal & Quality Challenges in AI-Generated News

• Copyright Disputes:
  • The NYT sues Microsoft and OpenAI for allegedly using its journalism to train large language models ([08:01]).
  • Chris Stoker Walker: “The outcome…could significantly affect the future of AI development. Because if the judge rules against OpenAI and Microsoft, then…large parts of what we believe might be in its knowledge base will have to be erased.” ([08:46])
• Quality and Accuracy Issues:
  • Even major companies like Apple suspended AI news summaries after persistent factual errors ([09:07]).
  • Chris Stoker Walker: “AI can be convincing even when it's wrong.” ([09:26])
• Newsroom Standards:
  • Many organizations (BBC, Guardian, AP) now have formal guidelines to oversee AI’s use in journalism ([11:17]).

4. Human-AI Collaboration at Telegraph Media Group

• Dylan Jacks:
  • The Telegraph has leveraged AI for years, focusing in recent times on generative AI’s potential to enhance, not replace, journalistic output ([12:06]).
  • Novel project: Using AI to translate and recreate podcasts ("Ukraine: The Latest") in other languages while preserving presenters’ voices ([12:35])
    • “To be clear, this is AI helping to present our journalism, not produce it.” ([12:47])
  • Strict editorial standards remain: “Our policy is really that…ultimately the person that's publishing that and their name's against that…they need to review it.” ([13:30])
• Audience Attitudes & Trust:
  • Trust and truth remain paramount; technology cannot outweigh the product’s integrity ([14:20]).
  • “There is only so much an AI can do. And being on the frontline of a conflict like Ukraine…is a big part of the value and how we communicate that.” ([14:20])

5. Risks: AI Disinformation and Manipulation

• Generative AI as Weapon:
  • Chris Stoker Walker:
    • Early fears about deepfakes have given way to concerns about data poisoning — manipulating training data to bias AI outputs ([15:36]).
    • “You can poison these models and therefore their outputs at source.” ([15:36])
  • Example: A Russian-backed network creates fake news sites to influence AI chatbot outputs ([16:11]).
    • “You can shift the balance of these AI models in a way that serves your goals rather than the goals of truth, reality and impartiality.” ([16:24])

6. Countering Disinformation: Fighting Fire with Fire

• Akin Tunde Bapatunde (Center for Journalism Innovation Development/Dubwa):
  • His organization has debunked over 27 AI-generated fakes in 2025 alone ([17:12]).
  • Common examples: Viral deepfakes of world leaders, altered celebrity images, and entirely fabricated incidents ([17:36]).
  • Disinformation can profoundly affect both health choices and democratic processes ([18:04]).
  • Innovative Tools:
    • AI model transcribes and flags questionable audio from radio, helping fact-checkers ([18:30]).
    • A free WhatsApp chatbot allows anyone to verify rumors in real time ([18:30]).
    • “You cannot fight fire without understanding how fire bites and how fire works.” ([19:20])

---

Notable Quotes and Memorable Moments

• On the initial wow factor of AI news avatars:
  • Adam Mossam ([04:49]):

    “It was probably the first time a lot of people saw some type of digital avatar with that much precision or that lifelike.”

• On why AI should amplify, not replace:
  • Adam Mossam ([05:21]):

    “Our product essentially acts as an assistant to the creatives. Right. It's really bolstering your existing staff and helping them get a lot more done in the same amount of time.”

• On the quality imperative:
  • Dylan Jacks ([13:30]):

    “Our policy is really that it's fine to use tools to help accelerate and draw together things, but ultimately the person that's publishing that and their name's against that, that is a representation of their work. They need to be confident and comfortable with that and they, they need to review it.”

• On data poisoning as the real disinformation risk:
  • Chris Stoker Walker ([15:36]):

    “Actually the real risk from disinformation through generative AI has been a much more subtle but no less pernicious concern, which is you can poison these models and therefore their outputs at source.”

• On using AI to combat AI-created fakes:
  • Akin Tunde Bapatunde ([19:20]):

    “You cannot fight fire without understanding how fire bites and how fire works.”

• On trust, truth, and journalism’s value:
  • Dylan Jacks ([14:20]):

    "...in news in particular, trust and truth...that is just paramount and the raisin d’être to why we're here..."

---

Timestamps for Key Segments

• AI Anchors Go Viral — [01:22] to [04:24]
• Channel One’s Approach: Human-free News — [04:16] to [06:33]
• AI’s True Role: Assistant, Not Replacement — [05:21] to [07:15]
• How AI Is Integrated in Major Outlets — [07:27] to [08:42]
• Legal Battles & Factual Errors in AI Summaries — [08:01] to [09:26]
• BBC, Guardian, AP: Guidelines for AI Use — [11:17]
• Telegraph Media Group’s Innovations & Boundaries — [12:06] to [14:20]
• Public Attitudes: The Trust Factor — [14:20] to [15:08]
• Disinformation Threats & Data Poisoning — [15:23] to [16:57]
• Fighting AI Lies with AI Tools — [17:04] to [19:26]

---

Tone & Style

The episode keeps a balanced and forward-looking tone — alternating between curiosity regarding technological advances and cautious skepticism about AI’s pitfalls.

---

Conclusion

This episode offers an in-depth look at the potential, limitations, and risks of AI-generated news. While technologies are rapidly advancing, widespread adoption will hinge on legal clarity, editorial guidance, and above all, audience trust. Both industry leaders and fact-checkers agree: AI is a valuable tool in the newsroom — but not a replacement for human judgment, ethics, and storytelling.

Transcript

[Chris Stoker Walker]

This BBC podcast is supported by ads outside the uk.

[Sam Grouet]

I love ravioli.

[Adam Mossam]

Since when do you speak Italian?

[Dylan Jacks]

Since we partnered with SAP Concur. Their integrated travel and expense platform and breakthrough solutions with AI gave me time back to dive into our financial future. We expand into Europe in 2027, so I'm getting ready.

[Adam Mossam]

Well, you can predict the future.

[Dylan Jacks]

I can predict you'll like that message.

[Adam Mossam]

What message?

[Dylan Jacks]

Oh, hey, we all got bonuses.

You can save for college now.

I don't have kids.

[Sam Grouet]

You don't say SAP Concur helps your business move forward faster.

[Chip Planoxel]

Learn [e-mail] the imperative for businesses has never been clearer.

[Adam Mossam]

The age of experimentation is over.

[Chip Planoxel]

We're talking transformation and winning at scale from AI that actually drives roi.

[Adam Mossam]

We are going through that curve of understanding what the technology really can and.

[Chip Planoxel]

Cannot do to turning reams of data into real competitive advantage.

[Akin Tunde Bapatunde]

A lot of these successful companies, they treat data as a product.

[Chip Planoxel]

I'm Chip Planoxel, host of Resilient Edge, the Smart Executive's guide to Implementing and sustaining Change. Paid and presented by Deloitte. Available now wherever you listen to podcasts.

[Sam Grouet]

Hello and welcome to Business daily from the BBC World Service with me, Sam Grouet. In December 2023, a video was posted on the social media platform X that drew a lot of attention.

[Adam Mossam]

We did, I think it was a little over 5 million views in the first 24 hours across several accounts, it.

[Sam Grouet]

Showed a team of AI generated news anchors flawlessly delivering the day's top headlines.

[AI News Anchor]

Military operations against Hamas by Israeli Defense Forces continue.

[Sam Grouet]

In the same year, media publication India Today launched its own AI anchor.

[Dylan Jacks]

She is our first AI anchor. Sana Mary on the job learning.

[Sam Grouet]

But could AI generated content ever become the new normal for news? All week on Business Daily, we've been exploring how artificial intelligence is changing the business world. Today, it's the turn of the news as we investigate how the news media is embracing the technology, the issues facing editors and readers alike, and how people around the world are fighting AI disinformation. That's all coming up on today's Business Daily. So I've been a journalist with the BBC for the best part of the last decade and what I'm about to do isn't part of my typical day. It's not part of my usual workflow. I'm going to use the most popular AI chatbot, that's Chat GPT, although others are of course available. And I'm going to ask it to write me the opening of a script on the use of AI within news media. So I want it to be conversational, and then I'm going to ask another bit of software, elevenlabs, to deliver it in a news reporter's voice. Okay, let's see how this goes.

[Dylan Jacks]

Artificial intelligence is already writing headlines, summarizing articles, even reading the news aloud to you. Some media companies say it's helping them work faster, cut costs, and reach more people. Others fear it could erode trust in journalism altogether.

[Sam Grouet]

Well, it's not perfect, and I might need to fact check some of the claims, but it's pretty impressive, especially considering it was made in about two minutes. And with software continuously improving and becoming more advanced, it begs the question, what could the future hold?

[AI News Anchor]

Hello and welcome to Channel One. A new way of consuming reporting and thinking about the news, powered by artificial intelligence.

[Sam Grouet]

For Channel One, the future is very much here.

[AI News Anchor]

All presented by our team of AI generated reporters.

[Sam Grouet]

The generative AI news service, launched for a brief period last year, advertised its news scripted, edited, and presented by artificial intelligence. Available in more than 30 languages and able to broadcast news stories in just minutes. But would you or anyone you know, actually watch a channel that lacks a real human being?

[Adam Mossam]

We did. I think it was a little over 5 million views in the first 24 hours across several accounts.

[Sam Grouet]

I put that to founder and CEO Adam Mossam.

[Adam Mossam]

But what was more interesting was we got, I think it was 8,000 emails in the week after that. And that was from such a wide variety of folks. Everything from people that were the creatives in the industry to the CEOs of. I feel like most of the biggest media companies on the planet trying to figure out what exactly that was and how exactly we did it.

[Sam Grouet]

And what do you think it was that captured everyone's imagination so much?

[Adam Mossam]

Well, I think we're operating on the leading edge of AI, and it was probably the first time a lot of people saw some type of digital avatar with that much precision or that lifelike.

[Sam Grouet]

Well, with many newsrooms making cuts due to changing budgets, Business Insider just laid.

[Chip Planoxel]

Off 21% of its staff, over one.

[Sam Grouet]

Fifth of the newsroom, while simultaneously announcing.

[Chip Planoxel]

That the news outlet wants to lean.

[Sam Grouet]

More heavily on artificial intelligence. What about the potential for AI to completely replace journalists?

[Adam Mossam]

Listen, there's a lot of jobs that, you know, if you speak to folks in the industry that are being done, that people don't really want to do. Right. They want to focus on higher value content. If we could take that off their. Off their plates. Our product essentially acts as an assistant to the creatives. Right. It's really Bolstering your existing staff and helping them get a lot more done in the same amount of time.

[Sam Grouet]

Adam, when you launched Channel One, you did it with the intention of having a news channel that was made completely using generative AI.

[AI News Anchor]

Next comes the launch of our daily news program in countries and languages across the globe.

[Sam Grouet]

Is that still the aim here?

[Adam Mossam]

Yeah, that's, that's definitely on the table for us. You know, the. To be quite honest, we made what we wanted to make and I didn't really like it. You know, me at my age and my team and our habits, we're not the type of folks that have the habit or the pattern of watching a 30 minute or a 60 minute news broadcast with 4 minute stories at a time. And we've decided that as a team, because we are maybe more mobile native or more social native, that we were going to focus our efforts there. And we've got some very interesting things that we're working on.

[Sam Grouet]

But will anyone actually watch a news channel presented by fake humans? Are you seeing the consumer demand for AI generated news content?

[Adam Mossam]

Yeah, I think, you know, to be honest, it's starting at the business to business level on the B2B side because you're seeing folks in these companies that just have ambition to do projects and that this were not possible a number of months ago. So we firmly, you know, from everything we've ever done with people, it has never been about staff reduction. It is firmly in the realm of what was not possible before. And what are the new possibilities we can enable for you?

[Sam Grouet]

So with new possibilities for news output, how much of it are we seeing rolled out across newsrooms around the world? Chris Stoker Walker is a journalist and author of How AI Ate the World.

[Chris Stoker Walker]

We are seeing AI being used on the front lines of journalism at the point at which the audience consumes that bit of information. So if you log on to Bloomberg, for instance, at the top of every story you will see an AI generated summary designed to take a 400 or an 800 word article and to condense it into three or four bullet points that you can very easily understand and take away.

[Sam Grouet]

But it's not all been smooth sailing between the media and AI.

[AI News Anchor]

The New York Times is suing Microsoft and chat GPT maker Open AI over copyright infringement. The newspaper claims the company's used its intellectual property to train large language models.

[Chris Stoker Walker]

It also one of the challenges here is that, you know, the generative AI space has been a little bit of a wild west. There is this real appetite for, for these AI models to be trained on data without that source information. AI needs that base data. And the New York Times obviously is one of the world's biggest publications, one of the most respected on the planet, is therefore a good source of ground truth data.

[Sam Grouet]

OpenAI has denied these claims and the case is ongoing.

[Chris Stoker Walker]

And frankly, the outcome of it could really significantly affect the future of AI development. Because if the judge rules against OpenAI and Microsoft Microsoft, then effectively large parts of what we believe might be in its knowledge base will have to be erased. And nobody's really understood quite how that would work.

[Sam Grouet]

Then in January this year, Apple was forced to suspend a new artificial intelligence feature after frequent mistakes were spotted in summaries of news headlines.

[Chris Stoker Walker]

Yeah, one of the things I think has been really interesting is seeing that actually the BBC itself is on the front line of this sort of dispute.

[Dylan Jacks]

Dispute.

[Chris Stoker Walker]

And the BBC has complained repeatedly to Apple about this. Of those Apple AI summaries of BBC news stories being factually incorrect, either misinterpreting the content of the actual news story or just making things up that were never in the news story in the first place. And the BBC has done an audit, effectively, of not just Apple intelligence, but of lots of these AI tools that it published earlier this year, which suggested that there are a worrying number of these errors that sneak through. And it comes down to that idea that AI can be convincing even when it's wrong.

[Sam Grouet]

You're listening to Business Daily on the BBC World Service.

[Dylan Jacks]

Hey, Ryan, that was a fast trip. It was like you teleported.

[Sam Grouet]

Yeah, just got in. I'll get all my expenses logged, I promise.

[Dylan Jacks]

Oh, no, you're okay. SAP Concur uses advanced AI, so your expense report will practically exactly right itself. Quite the breakthrough. It's like we've been teleported into the future. All right, so just curious, would you give us written permission to convert your matter into energy patterns and reassemble you at, say, random travel destinations?

[Sam Grouet]

Margaret, are you building a teleporter?

[Dylan Jacks]

No. Yes.

[Sam Grouet]

SAP Concur helps your business move forward faster.

[Chip Planoxel]

Learn [e-mail] We've had seismic changes in the last decade and we've adapted to them. But what comes next?

[Sam Grouet]

We really talk about autonomous business processes rather than just automating business processes.

[Chip Planoxel]

I'm Chip Clinicsel, host of Resilient Edge, a business vitality podcast paid and presented by Deloitte. Our latest episode explores how to build a digital brain and future proof your business vision to value on Resilient Edge is available now wherever you listen to podcasts.

[Sam Grouet]

I'm Sam Grouet Today, could artificial intelligence ever replace the news? Despite issues with generative AI, many publications, including here at the BBC, have recognized the need to implement it into news operations. And publications from the Guardian to Associated Press have drawn up guidelines for journalists. Let's go back to our AI generated news report quarter and I'm going to ask it to write me an introduction for my next guest. He's called Dylan Jacks and he's Group Technology Director at the Telegraph Media Group.

[Dylan Jacks]

Our next guest is at the forefront of how one of Britain's biggest newsrooms is using technology to shape the way we consume journalism.

So we are, I would say, eight years into how we use AI and targeting in the world of journalism, journalism and publishing. But we are probably two years into a much more focused AI era in respect to the emergence of Gen AI. But we really have been trying to position ourselves as an organization that drills into the value of how AI can support, elevate and accelerate journalism, including using.

[Sam Grouet]

AI for making audio content like podcasts.

[Dylan Jacks]

We've got a really interesting AI project around our Ukraine. The latest podcasts from today, each episode of Ukraine, the latest will be translated into, into Ukrainian.

[Sam Grouet]

We want to ensure those most affected.

[Dylan Jacks]

By the war can access our coverage. That podcast has got really great global reach, but one of the things that we wanted to do is we wanted to make that available to non English speakers, particularly in Ukraine and Russia. And we used very advanced AI tooling to recreate that podcast, cloning and using the voices of those presenters. To be clear, this is AI helping to present our journalism, not produce it. So it really preserved the sort of passion and the actual unique production of that, but made it available in that language.

[Sam Grouet]

But although keen to experiment, Dylan says the paper still has its boundaries.

[Dylan Jacks]

We want to make sure that the quality does not suffer at any point. So our policy is really that it's fine to use tools to help accelerate and draw together things, but ultimately the person that's publishing that and their names against that, that is a representation of their work. They need to be confident and comfortable with that and they, they need to review it.

[Adam Mossam]

Everyone is hating on Duolingo for going AI first.

[Sam Grouet]

Well, I dug up earlier this year, Duolingo became the latest big firm to come under fire for using artificial intelligence after the language learning app said it was going AI first.

[AI News Anchor]

And it's safe to say that the.

[Sam Grouet]

Public is not too happy about that.

[Akin Tunde Bapatunde]

The backlash on social media is so strong.

[Sam Grouet]

Across TikTok and other social media platforms, users said they were boycotting the app. So does Dylan think readers and viewers are ready for news made with AI?

[Dylan Jacks]

Well, I think that in news in particular trust and truth, I think, I think that is just paramount and the raisin destre to why we're here and getting too excited about the technology and not thinking about whether or not this is right for what we're producing. It's incumbent upon us to be progressive moving forward and looking at how AI can really supercharge our whole operation, but it obviously can't undermine the product itself. And there is only so much an AI can do. And being on the front line of a conflict like Ukraine and really getting to grips with the sort of stories and the people affected is, you know, is. Is a big part of the value and how we communicate that that is what our readers are interested in and that's what, that's what, that's the value that we can add.

[Sam Grouet]

While Dylan admits AI has its limits when it comes to journalism, there are others that are concerned about its use in the spread of disinformation, particularly when that's done with the specific aim of undermining democracies around the world.

[Chris Stoker Walker]

Disinformation is a real risk, I think with generative AI for not necessarily the reason that we initially thought.

[Sam Grouet]

His journalist and author of How AI Ate the World, Chris Stoker Walker.

[Chris Stoker Walker]

There were initially some concerns that generative AI might be used to create, for instance, deep faked videos of world leaders either in compromising situations or saying things that they might not want to. We did actually see early examples of this in the Russia Ukraine war. But actually the real risk from disinformation through generative AI has been a much more subtle but no less pernicious concern, which is you can poison these models and therefore their outputs at source.

[Sam Grouet]

A study by NewsGuard, a company tracking online misinformation, found that a Russian network called Pravda is creating news websites to influence the answers given by AI chatbots.

[Chris Stoker Walker]

So there has been uncovered a kind of Russian backed state sponsored attempt to try and pollute the training data that these AI models rely on by feeding narratives that are more in line with Russian sensibilities. It does highlight just how easily you can shift the balance of these AI models in a way that serves your goals rather than the goals of truth, reality and impartiality.

[Akin Tunde Bapatunde]

My name is Akin Tunde Bapatunde. I'm the executive director of the center for Journalism Innovation Development.

[Sam Grouet]

It's responsible for Dubwa, a fact checking organization fighting disinformation across 13 African countries.

[Akin Tunde Bapatunde]

For Me, I've worked on this information for about eight years now. Between January this year and now, we've actually had to debunk about 27 AI generated fakes, which is a big push from what used to happen.

[Sam Grouet]

So are you able to give some examples of fake stories you've seen?

[Akin Tunde Bapatunde]

Right, we've seen a viral video of President Trump criticizing the Ukraine president. Right, we've seen that. You know, we've seen that in Nigeria, we've seen pictures of popular Nigerian child comedian doctored in Sierra Leone. In Liberia, we've seen video of crashed plane in muddy water as fake.

[Sam Grouet]

And what's the danger of reading so called news that isn't actually truthful at all?

[Akin Tunde Bapatunde]

We become who we become as a result of what we consume as information. We're conducting a research that's looking at how disinformation impact people's health choices. Oh, there's election and then you come across disinformation and that impacts the kind of people you choose as your leader.

[Sam Grouet]

I ask Akintende how Dubuis is fighting this disinformation.

[Akin Tunde Bapatunde]

There's a lot of disinformation in audio formats on radio stations. So we've built an AI model that could transcribe audio to text and then it would automatically flag potential disinformation that can point fact checkers to what they can do. We've also built a chatbot that's easy for anybody to chat with. It's on WhatsApp and it's free. If there's a viral content that you need to know whether it's been debunked or not, just chat with the bot. And then when you chat, you could verify a claim and it would aggregate evidence of credible sources to debunk whatever rumor is ongoing.

[Sam Grouet]

It's amazing to hear about the work you're doing and you're actually fighting fire with fire. So there's a problem with generative AI disinformation. You're using Gen AI to fight back against it.

[Akin Tunde Bapatunde]

Right. That's a paradox, isn't it? You cannot fight fire without understanding how fire bites and how fire works.

[Sam Grouet]

Thank you to Akintunde and to all of my guests. You've been listening to Business Daily on the BBC World Service with me, Sam Gruy. This episode was produced by Megan Lawton. Tomorrow, Hannah Malayan will be speaking to businesses and individuals choosing not to use AI and asking whether they'll be left behind as a result. You can find more episodes just search Business Daily. Wherever you get your BBC podcasts.

[Dylan Jacks]

This.

[AI News Anchor]

Is the story of the one. As a custodial supervisor at a high school, he knows that during cold and flu season, germs spread fast. It's why he partners with Granger to stay fully stocked on the products and supplies he needs, from tissues to disinfectants to floor scrubbers, all so that he can help students, staff and teachers stay healthy and focused. Call 1-800-granger click grainger.com or just stop by Granger for the ones who get it done.