The Washington Post Senior Editor for AI Strategy and Innovation Phoebe Connelly shares some key things she has learnt and explains why she’s excited about the future of journalism. Lucinda Jordaan conducted the interview and wrote this article
The convergence of AI and journalism in the second digital transformation calls for new newsroom applications, processes, and roles. The Washington Post has been progressively experimenting with AI and last month, it launched Climate Answers, its first AI chatbot. Phoebe Connelly is the journalist at the forefront of The Post’s tech-editorial convergence. “I think one of the most exciting parts about being in journalism right now is that there’s a good chance the job you’ll have five years from now doesn’t yet exist,” she says.
In February, Connelly became the title’s first ever senior editor for AI Strategy and Innovation. It’s not her first ‘first’: as director of The Post’s Next Generation Audiences, she led an experimental team that boosted acquisition of younger and more diverse audiences through new products, practices and partnerships. Interactive storytelling is her forté: Connelly joined The Post in 2013 after serving as senior news editor at Yahoo! News. As deputy director of video, her team was awarded the Pulitzer Prize for Public Service for its coverage of the 6 January insurrection at the US Capitol. The Post also won Digiday’s Video Team of the Year Award in 2018.
Phoebe Connelly speaks to the World Editors Forum about the exciting possibilities for the future of journalism in a new media landscape:
What exactly does your role entail?
The role changes day to day, I think, as with any role. But I am meant to straddle two worlds: I report to both the editor-in-chief (Matt Murray) and the chief technology officer (Vineet Khosla). We’re approaching AI as we approach every new technology, by asking: ‘What’s the nimble, smart set of stakeholders we can pull together to help us conceive of ways to serve our readers better?’
Sometimes that will take the form of a brainstorming meeting with key members of the newsroom and technologists from the engineering side, and we hash out what’s possible; what the user needs are, what resources and idea we already have that we can marry with a large language model (LLM) to produce something fresh and inventive.
Sometimes it’s in more of an advisory or consulting role, pulling together key newsroom stakeholders, our standards team, and making sure that we craft thoughtful policies for how we’re approaching the use of AI. Sometimes it’s just meeting with a journalist and talking through use cases that they might think about for tinkering with AI on their beat in a way that aligns with our standards, but pushes us forward in terms of tools and techniques.
Did you have AI on your vision board or your five-year game plan?
I didn’t have AI on my vision board. But I’ll say this: I think one of the most exciting parts about being in journalism right now is that there’s a good chance that the job you’ll have five years from now doesn’t yet exist. And that can be truly terrifying, because I think we all like certainty in thinking about our futures – but my current job and my previous job didn’t exist five years back on my board. What was on my vision board then was tackling new technology, trying to reconcile it with all of the strengths of our industry, getting to work with new teams across the organisation, and really engaging with the questions of how we can best serve our audiences.
What learnings from NextGen, are you taking into AI?
With the NextGen team I was asked to focus on our users, and figure out how we could better serve those users, and how that fit into our daily practices of journalism. I think what I’ve brought with me isthat focus on the user, putting that user first. I think AI offers promise to our industry only insofar as we use it to tackle the needs of our users; whether those users are internal – our own reporters and editors who need access to AI to speed their reporting – or whether it’s a loyal subscriber, or a new reader whom we need to delight and inform and bring the best of our journalism to bear.’
What are you doing to help the newsroom adapt to this new tech – and what early learnings from this experience can you share with others?
The people who are going to be able to tell us how to use AI responsibly in newsrooms are the journalists themselves. The most useful approaches will come from the departments, the journalists themselves who are really thinking structurally about how to deliver the best of their beats, down to the youngest, newest reporter. In my role, I can convene the right conversations. I believe that we will move forward as an industry by assuming that the solutions and the ideas about how we use AI come from all levels of the company, as opposed to from the top down, or just from the engineering side, or just from the newsroom side.
I think the degree to which we set aside and encourage ownership to be shared with the teams that are going to use these tools, is how we’re going to increase adoption and really build something strong. In a breaking news situation, we are comfortable with sending ad hoc teams of reporters and photographers, videographers to the scene, to work together to produce great storytelling.
That’s the approach we are taking in building products or adopting new technologies at The Washington Post: pair teams of reporters, editors, product people, and let them iterate quickly; give them guidance and see what they come back to us with. And that’s where it gets exciting: we recently launched Climate Answers, and it almost instantly sparked ideas with other journalists in the newsroom. So right now, I’m in an intake mode – you’ve caught me in a week where I’ve already met with three different desks, who have ideas on how we can take that forward, and get the next project going.
What advice to other newsroom editors, who are obviously mostly overwhelmed by this new tech, can you impart?
As you’re shaping your approach and your policy, you should be talking to the most enthusiastic people – as well as the people who doubt it the most. That’s how you’re going to come up with a strong approach, by making sure you’re hearing from both ends of the spectrum – and every newsroom has that. Secondly, start experimenting, and start looking to see where others, non-journalists, are experimenting. The best idea may not come from other journalists – it may come from somebody in an academic space, a company or an individual. Our job is to make sure that we’re doing it in service to and in alignment with our principles, but we should feel open to inspiration coming from any place.
Where are you getting your inspiration from; what are you reading right now?
This is the summer of Matisse for me; I’m in the first of a two-volume biography on Henri Matisse, the painter. I’m finding it inspirational in all sorts of ways: how he approached visuals; the shift he made in colour palette, in form, in subject… We think of him as this fixed, expert painter, but he was constantly on this learning journey, and by the end of his life he built this beautiful chapel in the south of France. I find the ways in which he stayed open and free, and kept challenging himself, hugely inspirational. So that’s my non-AI read that is informing how I approach tools in an open way.
I’ve also just read a great interview with philosopher Dr Ruth Chang, on decision making, and her thoughts on how we should approach AI systems – and an excellent NiemanLab piece: Andrew Deck’s interview with filmmaker Errol Morris. Right now, I’m really interested in learning from people who are not invested in AI. Of course, we need to talk with AI and LLM experts, but I find it revelatory to also talk with experts outside the field, in thinking about how we use these tools.
(By special arrangement with WAN-IFRA. Lucinda Jordan freelances as writer, editor, consultant and coach: providing full-suite media and communications services to media enterprises and agencies. She regularly writes for the WAN-IFRA World Editors Forum.)
from Webdoux