I have spent the past couple of months talking to dozens of INMA members about GenAI use cases. 

The vast majority of them are uses for editorial, perhaps because that is where most companies expect to find efficiencies that can really move the needle.

Why? Because newsrooms are a substantial part of the business for any media company, and GenAI has shown that it generally works well as an editorial assistant — e.g., transcribing and summarising articles, tagging them with metadata, or coming up with story ideas or headlines that an editor then vets.

But one conversation repeatedly keeps popping up: What about uses for GenAI in advertising? Surely it has tremendous potential to generate higher revenue for our business, especially as we worry about the end of third-party cookies.

Today’s newsletter takes a look at two large, highly regarded news publishers that are experimenting with deploying the technology for different uses in advertising with an eye on charging premiums from advertisers. 

We also dip into 11 other GenAI use cases that are useful for media companies.

GenAI for advertising

A couple of INMA members have told me that they are experimenting with using GenAI to come up with ideas on the creative side of marketing and advertising.

But there are two particularly interesting developments on the advertising revenue front, one from a legendary U.S. newspaper and the other from an innovative Indian publisher that is also an institution.

The first is The New York Times, which plans to use GenAI for real-time ad targeting and optimisation. 

The news publisher is building technology that will recommend where an ad campaign will perform best, based on its content. It would let The Times align the ad’s message according to where the ad is placed, allowing it to target different niche audiences. 

“This experimental product uses Generative AI to automatically align a brand or a specific campaign’s message with relevant editorial content, identifying ideal contextual environments that are unique to The Times and expanding beyond what is traditionally available for contextual targeting options,” a Times spokesperson told me.

“As a result, our new product will find audiences who are most interested in certain articles and will target them anywhere else on The Times, expanding the campaign’s reach and adding nuance to its audience targeting.”

The advertiser no longer will need to choose targeting criteria before the campaign is actually deployed — they can be chosen in real time. The technology also ensures ads will not appear adjacent to unsuitable content.

The Times is now finding advertisers that would like to beta test this technology in the second quarter of the year. Pricing will be determined later. The Times has been working on this technology for the past nine months.

The second example is from Delhi-based Times Internet, India’s biggest digital products company, which owns the largest-selling English daily in the world, The Times of India. The publisher is experimenting with an advertising chat product that interacts with the user and keeps them engaged for longer — an innovation that could allow them to charge a premium from advertisers and create sustained value for them.

Picture a banner ad on a Web site for, say, an electric vehicle. Now, imagine if you clicked on it and a chat box opened. You can ask questions about the vehicle here — e.g., how much is the range of the car, can I get it with leather seats, etc.

The company started building the product a few months ago and is still fine-tuning it.

Are you experimenting with GenAI for advertising? Please let me know!

11 GenAI uses for media companies

Here are some practical applications of GenAI that have emerged over the past couple of weeks:

1. GenAI for chat products: How can you learn from others’ mistakes and get a chat product right? Watch a recording of our Webinar (or read our summary) that saves you from reinventing the wheel. You can also watch a recording of a Webinar featuring India’s Jagran New Media (blog summary here), which has also built a chat product for search that its users find highly engaging.

2. GenAI investigative journalism tool: A candid piece by reporter Jaemark Tordecilla on how it took him 16 hours to build a truly helpful tool that he expects will save reporters 80% of the time they spend on audit reports, combing through data.

3. GenAI for diversity: The Baltimore Times has created a set of inclusive avatars and voices that allow its audiences to select from multiple diverse “personas” that reflect and represent the audiences they seek to serve.

4. GenAI for video: I have seen mainly rave reviews for OpenAI’s new video tool Sora, which can create extremely realistic, high-quality details. It does raise questions about copyright and training data, though.

5. GenAI for classifying information: Semafor found it was hard to collect hate-crime statistics because they are defined and tracked in different ways by local U.S. police departments. So, Executive Editor Gina Chua plugged a definition of hate crimes into a bot and was surprised by how well it could discover unwritten rules and relationships in articles. Chua suggests this capability could be used for “content moderation or finding violations of a given policy in a sea of complaints.”

6. GenAI for rewriting copy: Reach Plc uses it to repurpose copy from wires, police releases, and elsewhere in its own chain (as do Newsquest and Ippen), as well as image captions. 

7. GenAI for marketing: WPP unveils text and image-creation tools and talks about training models on brands’ tone/voice or target demographics. So far, they have almost 30,000 users and millions of prompts. It is part of a broader trend of using AI in marketing — design agency Code and Theory has joined a creative AI partnership with Oracle, Publicis has fully acquired Publicis Sapient AI Labs, and Omnicom has joined forces with Microsoft. 

8. GenAI for PDFs: Adobe makes a PDF reader chatbot available in Acrobat and Reader in beta. The road map includes integrations with Firefly, the ability to pull information from multiple documents, document types, and sources simultaneously, and features for generating first drafts and editing copy.

9. GenAI for tapping into new audiences: Schibsted gets in younger listeners’ ears with audio reports and finds high completion rates and greater engagement than for text articles. Jagran is building an AI-based app for younger audiences.

10. GenAI as an alternate source of revenue: Reddit is reported to be selling its users’ conversations to Google as training data for US$60 million ahead of an IPO. 

11. GenAI for customer service: A cautionary tale for anyone experimenting with this use case. The bot can make up stuff, and you are on the hook for it.

What I’m hearing

Idea generation: A lot of great GenAI use cases seem to have come out of hackathons. Are you running them at your company? 

• Ringier Axel Springer mentioned its “huge hackathon” as the reason for creating a secure environment for employees to play with GenAI. 

• Spread the net wide: Jagran collaborated with India’s top engineering college for its hackathon. 

• And hear from Brazil’s Grupo RBS on how it is working on some great experiments that came out of its hackathon on our next GenAI Webinar.

Worthwhile links 

• Productivity gains? They will not be straightforward, says this report.

• ROI? It’s not straightforward either at this point.

• Carpe diem: With GenAI, let's not repeat the mistake we made by dragging our feet on technological change, says the Aspen Institute.

• ICYMI: A thoughtful report on AI in the news business by Felix Simon at the Oxford Internet Institute, highlighting the growing power of tech companies.

• Google upgrades ad tech and enterprise tools by incorporating Gemini to compete with OpenAI. 

• No, thank you: Around half of the top news sites in 10 countries block AI web crawlers.

• Transformation: Don’t just throw AI tools at people and expect a transformation. Think about workflows instead.

• The AI search revolution: Read and, well, don’t weep — get going on a solution instead: The dawn of personalised Web pages powered by AI search.

Date for the diary: March 6  

An INMA Webinar on insights into using GenAI will feature two smart techniques: Prompt engineering for a print-focused chain, and gathering audience feedback and turning that into journalistic coverage. Don’t miss our Webinar on Wednesday, March 6 at 4 pm CET/10 am ET.

An AI diversion 

My Spotify playlists are the objects of derision — first, from my family and now from a snarky bot. Go on, try it and see if you get more respect than I did.

About this newsletter

Today’s newsletter is written by Sonali Verma, based in Toronto, and lead for the INMA Generative AI Initiative. Sonali will share research, case studies, and thought leadership on the topic of generative AI and how it relates to all areas of news media.

This newsletter is a public face of the Generative AI Initiative by INMA, outlined here. E-mail Sonali at sonali.verma@inma.org or connect with her on INMA’s Slack channel with thoughts, suggestions, and questions.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .