SIIA joined a multiassociation letter to request further consultation on the next draft of the proposed Telecommunication Bill, 2022 (Draft Bill), first circulated for public comment on 21 September 2021. You can read the full letter here.
SIIA Submission to the Office of Science and Technology Policy on National AI Strategy
SIIA wrote a submission in response to the request for information on National Priorities for Artificial Intelligence (RFI) issued by the Office of Science and Technology Policy (OSTP). SIIA appreciates OSTP’s continued attention to advancing responsible innovation in artificial intelligence (AI) and its efforts to develop a comprehensive, whole-of- society approach to harness the potential and mitigate the risks of AI technologies. You can read the full submission here.
Empowering Educators with Innovative AI-Based Assessment Solutions – A Look into Cambium Assessment, Incorporated
Cambium Assessment, Incorporated (CAI)is revolutionizing the field of educational assessment by empowering educators with innovative AI-based solutions. Their use of state-of-the-art artificial intelligence, combined with human scoring, enables efficient and accurate evaluation of student writing. CAI’s commitment to responsible AI use and their adherence to key principles in the Blueprint for an AI Bill of Rights, showcasing their impact on the educational landscape.
CAI uses state-of-the-art artificial intelligence (AI) to automatically score student writing with our Autoscore engine. CAI deploy their models to work alongside human scoring; unusual or borderline responses are routed to human scorers for verification scoring, and random subsets of responses are routed to human scorers as a quality check. Their approach helps to ensure that all responses receive valid and reliable scores and that scores are returned quickly to educators for use in the classroom.
CAI’s approach addresses each principle in the Blueprint for an AI Bill of Rights.
- Safe and Effective Systems.
- Algorithmic Discrimination Protections.
- Data Privacy.
- Notice and Explanation.
- Human Alternatives, Considerations, and Fallback
‘Think Ambitious Experiments’; Valuing Failure and Coaxing the Best Ideas From Colleagues
It’s interesting how we create our best work. I recall as a young Washington Post sports reporter being told by my editor that a tennis article wasn’t good enough, and then angrily writing something much, much better. John Kander, the composer of Cabaret who just passed at age 96, told a similar story about writing the wonderful song New York, New York with partner Fred Ebb, after Robert DeNiro told him their first attempt wasn’t good enough for the film. “What does he know?” Kander recalled asking.
But you can’t go to that well too often, as they say. From talking to a lot of industry folks, one thing becomes very clear: The importance of giving people the space to try, fail and try again. Here are some examples:
Normalize talking about failures. “Working with metrics is all about trial and error, adjustment and retrial,” Elizabeth Gamperl wrote in her Reuters Institute report. “Every failure is a step closer to success.” Said one editor: “We have as many open conversations about when things haven’t worked as possible without everyone getting really upset. That is not easy because people work incredibly hard in the newsroom. What lessons can be learned?”
Make failure safe. “You have to have the ability to put yourself out there and be willing to fail,” Heather Farley, CEO of Access Intelligence and a confirmed speaker at BIMS 2024, once told us. “Fail fast and fail forward is my favorite motto.” What was one of the first things you did when you became CEO of Match? Sam Yagan was asked. “When I took over Match, I realized that they use data, but the expectation—which was always data-driven—was that tests will all succeed. It wasn’t built in a culture of failure. I compare never failing with not having ambition. [So the question became,] how do we let ourselves test out our intuition? The intuition has to inform what data you get.”
Learn lessons in success and failure. “If something [messes] up, you can look at your stats and figure out what went wrong,” said Kate Lucey, a former digital editor for Cosmopolitan UK. “Try new ideas—if they work, how can you expand them? If they fail, why did they fail and what have you learnt about your audience that you can apply to future work? It’s constant learning, constant adapting—and a constant headache… but it’s FUN.”
“Create a culture to build trust and collaboration, and breaking down silos…” Tim Hartman, CEO of GovExec said at one of our conferences. “Think ambitious experiments and trust each other. If you look around and don’t see that, you have a problem.”
Don’t let the quiet ones stay quiet. “Have a think tank where you can bring people to brainstorm,” Elizabeth Petersen of Simplify Compliance told us once. “Every person has ideas but you need to coax them out. I like to brainstorm on the fly. I have introverts [on staff], and they need to be encouraged. To have a structured agenda is a great way to get people talking.”
Set benchmarks.“One of the biggest barriers to innovation is fear of failure,” Petersen added. “The information industry is changing so rapidly and there are so many unknowns. Even the most thoroughly researched product may not gain market traction. This is similar to the challenges faced by the best betting sites not on Gamstop, which operate outside traditional regulations and must navigate significant risks to attract and retain users. The key to developing a humming new product development engine is to be comfortable with risk and to set measurable (and transparent) benchmarks for product success.”
Allow for turbulence. “Embracing failure is easier said than done,” Anita Zielina, former director of news innovation and leadership at the Craig Newmark Graduate School of Journalism at CUNY, told us at BIMS a few years ago. “We like to win and are not so excited about failure. But the culture of failure empowers your team to experiment. If you don’t, you’re not going to have creativity in the room. Experimentation includes failure, and organizations need to live with that. There is no digital product development that doesn’t have unexpected turbulences. But it also allows for agility.”
‘Frustrate Them Just the Right Amount’; Publishers Search for the Perfect Paywall
Determining the right balance for your paywall is a never-ending—but thoroughly important—quest. What to put in front of the paywall? How much to put there? If metered, how many? For a while, INMA was doing timewalls, giving readers a day or so to read an article before it goes into subscriber-only mode. It was cool to watch the clock go down, but I’m not sure of its logic. Here’s some practical advice.
“The key here is to balance frustration and engagement,” said Madeleine White, head of international at Poool, a French start-up focused on audience conversion strategies, in an article on FIPP about Elle’s successful subscription model.
“Ultimately, this is what decides whether someone’s going to convert or not. You engage users with your newsletter, with valuable content, the range of UX features, and then you frustrate them just the right amount with a conversion wall to encourage them to give you some value in exchange for some other value.”
I asked one of my go-to industry savants, Matt Bailey of Site Logic and the Endless Coffee Cup podcast, how he handles paywalls. “For my training content, I have a number of ‘always free’ programs, and I keep adding to them,” he wrote me. “It makes a great hook and keeps its own audience. My regular courses have a bigger price on them, so it’s a big jump for some people. Having a semi-regular free session builds my mailing list and potential buyers.”
Bailey went on to assess some of what AI vendors are doing. “What I’ve seen lately—it’s highly effective to have access to a portion [of your program]. This is what’s happening with a lot of new AI programs right now. Free, but limited hours, requests, files, etc. in order to prove the ability of the program.
“I bought an annual subscription yesterday for an AI software that creates audiograms—because the free access limited me evaluating 30 minutes a month. Once I tried it, it was amazing and I bought it. I think there is a reason that most AI programs are using that hook—it works.”
Here is more advice about paywalls:
Don’t hide your premium content. A common problem for publishers, the FIPP story said, is that their premium content often goes unnoticed on their website. “On average, half of visitors will be lost at this stage in the funnel, with people not reaching the premium model or paywall. It’s crucial then that a brand’s subscription offer is more visible to move people through the funnel and get them to discover the value of the premium content.” During a recent industry workshop, experts drew an intriguing comparison to casinos not connected to Gamstop, emphasizing how these platforms excel at attracting and retaining users by making their offers impossible to ignore through strategic visibility and targeted messaging. ELLE applies a similar approach by giving premium articles a “sticky header with a CTA on it, adding a promotional banner and putting another promotion at the end of the story. A tag clearly identifies it as premium content, which is easily noticeable since everything is one color.”
Make it easy to find. According to Poool’s White, “52% of visitors to premium content never see the paywall because it’s too late to load, loads after adverts, or is just not high enough up the page.” At Elle, moving the paywall up the website by just 10% had an impact on conversion rates.
Don’t give too much away. Sarah Ebner, executive editor and head of newsletters for the Financial Times, “was surprised to see that [their] breaking news alerts consisted of a few paragraphs with a ‘read more’ CTA at the end. Putting so much detail in these emails meant that there was no huge impetus to click through,” she wrote. “We had given too much of the story away already. We made some simple changes, changing these alerts to only one paragraph, and altering ‘read more’ to ‘read the full story.’ The click-through rate went up by 41% in a month.”
Offer discounted trials. “An analysis of 35 leading news subscription organizations has shown that the price of digital subscriptions has fallen since 2017 in real terms—at least £2 per month once adjusted for inflation,” Tara Lajumoke, managing director of FT Strategies, wrote. “Of these publishers, 27 had some sort of heavily discounted trial. These tactics allow organizations to convert their non-core target audiences, without hurting the ARPU of their existing subscriber base.”
Get tougher. Most publishers are too generous and need to stop more readers to force conversion, a Shorenstein report said. You want to achieve a high stop rate—that is the percentage of all digital users who are “stopped” by a subscription prompt, a paywall or a meter limit. It is calculated by the number of users stopped by a meter or paywall in a given month over the number of unique visitors during that period.
Lower your meter limit. A majority of publishers with metered models set their meter limits at 5 articles per month or lower. This number has gone steadily down since 2012. Some publishers used to set the paywall as high as 25 articles a month. “As publishers have experimented, and readers have become accustomed to digital subscription, meter limits have tended to decline among the publishers studied and within the industry at large,” the Shorenstein report said.
Increase reader opportunities to encounter the meter. Is the meter simply the articles a reader clicks on, or are there more factors involved? You might lower the meter rate for more editorially-intensive content. Their limit might be increased if they do other things with you. For those with an ad blocker, a subscription message might be customized to invite the reader to subscribe or turn your ad blocker off to continue to read content before the average meter stop.
