Roa Powell

ippr.orgAI's got news for you: Can AI improve our information environment? | IPPRAlready, 24 per cent of people report using AI for information seeking every week. But there is widespread concern that the information provided will be in

AI is becoming a new front door to the news - 40% of 18-24 year olds use AI for information weekly and Google AI overviews reach 2 billion a month. But we find AI isn't reliably surfacing news content - with implications for public access to information and media plurality. Full report @ippr.org

Carsten Jung@carsjung.bsky.social

New report: AI is certain to transform the way we consume news. If we do nothing it could turn it into slop. We need policy action now, to ensure AI improves news & the public sphere. Guardian coverage: www.theguardian.com/technology/2...

theguardian.comAI-generated news should carry ‘nutrition’ labels, thinktank saysThe Institute for Public Policy Research also argues that tech companies must pay publishers for content they use

independent.co.ukGoogle faces £1 billion legal claim from UK consumers over Play Store feesClaim alleges that Google broke competition law by imposing a commission of up to 30 per cent on app developers for purchases made through its Play Store

Over £1bn claimed for UK consumers in today's Google Play trial, clearly showing the stakes for the public when digital competition is undermined and prices become uncompetitive. But class actions like this can only fix harms that have already taken place. 1/2

ippr.orgFair play: How competition policy can drive growth | IPPRAt the same time, the UK economy is suffering from chronic low investment, and firms have increasingly cited uncertainty around competition enforcement as

@ippr.org has repeatedly recommended bolder policies that prevent these sorts of excess charges in the first place - putting money back in the pockets of both consumers and UK innovators The CMA is currently deciding on rules to fix these app store payments and government must back a bold approach

Damning testimony from the CMA's former chair, chief economic advisor and cloud inquiry chair on how government has tackled big tech market power. Quotes from our @ippr.org report. Bold enforcement of digital competition policy is not anti-tech, and should be part of our agenda on homegrown AI.

Roa Powell reposted

Carsten Jung

@carsjung.bsky.social

Business dynamism is crucial for growth. But in digital markets, big tech behemoths are holding back UK businesses. Our new study shows that tech-intensive firms are not getting a good deal from big tech. Politicans need to empower the CMA to enforce competition properly! bsky.app/profile/roap...

Roa Powell@roapowell.bsky.social

The UK is spending £500 million to build British AI champions while leaving in place a market structure that will see them acquired or outcompeted. New report today on digital competition for growth and sovereignty - with polling of 417 UK businesses that rely on tech. Here's what they told us 1/6

Roa Powell reposted

Jack Stilgoe

@jackstilgoe.bsky.social

Should be possible to gather new coalitions around the diagnosis that big tech are anti-innovation

Big tech dominance is stopping new ideas getting off the ground. More than a third of firms we surveyed had shelved a product because of a major tech company dominating the market. Stronger competition means more room to innovate and grow. Find out more at www.ippr.org/articles/bottleneck-britain

92% support conduct rules the DMCCA already permits. And on average 72% say they would lower prices or improve products if the rules we tested were introduced. 5/6

ippr.orgBottleneck Britain: How digital competition could power UK growth and sovereignty | IPPRThese companies hold extraordinary control over the digital infrastructure we all rely on. They have earned much of their success and deserve to profit, bu

As a result we are recommending a new strategic steer for the CMA. Ministers need to back bold enforcement and recognise this could be part of their growth and tech sovereignty agenda.

Meanwhile resilience is suffering. 62% say switching from their main tech provider would be costly and disruptive. 7% couldn't switch at all. And in the mean time half of businesses would cease operating in days if they lost access. 3/6

Asked what constrains their growth, businesses ranked big tech market power above access to finance and talent. Regulatory costs ranked higher still, but competition isn't what they mean, as most businesses experience this as an enabler not red tape. 4/6

79% are concerned about dominant platforms limiting competition and the harms they experience are concrete. Over a third have shelved a product because a tech giant already dominated that space while close to half have seen platforms favour their own products. 2/6

open.substack.comWorkers actually have significant leverage right nowBusinesses are desperate for workers to adopt AI, but what offer are they making?

10 July

This week we have a guest post on our substack from @adamc-c.bsky.social Adam looks at why workers have unique leverage over AI adoption right now - and how we use that to shape deep AI transformation. Really recommend giving this one a read!

modeleconomy.substack.comOf course, AI has a politics. The question is, who is in control?Making AI neutral sounds like a fair goal, but if it is never fully possible, how can we manage the biases that remain?

2 July

I wrote something short on the recent discourse about AI being left wing. Nobody is ever going to agree on what "neutral" looks like for AI - just look at any debate about BBC neutrality. So, what can we practically do to ensure AI isn't acting as an opaque persuasion machine?

economist.comThe AI backlash is only getting startedHere is how to deal with it

26 June

Bullish optimism on AI no longer seems to be working

ippr.orgAcceleration is not a strategy: A framework for directing AI towards public value before it's too late | IPPRThe public are understandably worried about AI and, so far, governments have struggled to articulate a clear vision for what it would mean for AI to go wel

26 June

Good to see lots of our @ippr.org recommendations repeated here: 1. Redistribute AI's upsides, 2. Get better at measuring the economic impact and 3. Embark on more ambitious public good AI missions

open.substack.comopen.substack.com

23 June

“What should exist? To me, that’s the most exciting question imaginable. What do we need that we don’t have?” Positive visions of a future transformed by AI tend to come from Silicon Valley, but in this new Substack @carsjung.bsky.social explores how we could all have a say.

Carsten Jung@carsjung.bsky.social

New AI economics piece out today. As governments are understandably racing to get AI built out and adopted, one question seems missing: What do we actually want to build? open.substack.com/pub/modeleco...

modeleconomy.substack.comAI jobs predictions keep missing the point. Here’s what to measure insteadA government institute to track how AI disrupts jobs is long overdue. But as it officially launches today, will it measure the right things?

8 June

Some news… @carsjung.bsky.social and I are launching a Substack! We will be using this to explore what needs to change for AI to actually work for most people (spoiler: it is a lot) First post is up looking at today’s launch of the AI economics institute and how to predict AI's jobs impact 1/2

8 June

I take a look at the messy world of AI job loss predictions where the people building the technology oscillate between 'half of jobs gone in 5 years' and 'nothing to fear' An independent voice cutting through the noise is essential - but it won't be easy. So, where should the institute start? 2/2

1 June

Great report from @josephevans.bsky.social And it is not just @ippr.org recommending portable benefits to support workers through the AI transition - even OpenAI have floated the idea. If AI's biggest winners agree, what are we waiting for?

ippr.orgStrike while AI is hot: Rebuilding worker power for the age of AI | IPPRArtificial intelligence is the latest technology to raise the spectre of human labour becoming obsolete. AI technologies can now perform complex cognitive

🤖 | NEW REPORT: AI is reshaping work, and too often workers are paying the price. In this new paper, @josephevans.bsky.social sets out a new generation of worker rights so AI works for people, not just profits. Read more here 👇 www.ippr.org/articles/str...

Roa Powell reposted

Giles Wilkes

@gilesyb.bsky.social

22 May

IMHO possibly the most interesting question in finance and economics right now is the extent to which future AI dominance can retain any monopoly characteristics. People assume because it's "Silicon Valley" that those explosive network effects will necessarily pertain. They really might not... 1/

politicshome.comThe Consequences Of Inaction On AI-Driven Job Loss Are Coming Into ViewThe scale of the disruption to the labour market from AI is becoming clear; we cannot leave it to big tech or populists to frame the debate about w...

15 May

We won’t know the scale of AI’s jobs impact until late in the day. But if governments wait until then to prepare, state capacity will be grinding into gear just as people need support. I wrote for @politicshome.bsky.social on what government should do now: www.politicshome.com/news/article... 1/3

15 May

AI companies are calling for policies which on the surface feel progressive - taxes on their profits, redistributive wealth funds and strong worker power. When even AI’s biggest winners are pushing for redistribution, it is time to government to be thinking big. 2/3

15 May

But we can’t let the companies lead the debate. AI’s impact is already uneven - 60% of high earners use AI daily vs 16% of low earners and young people are facing the most urgent impacts Government needs a clear voice on this, before backlash escalates and only the simplistic answers are left 3/3

politicshome.comWe're Going To Need A Bigger Stick: Britain Seeks AI SovereigntyBritain has now made clear that it wants AI sovereignty, of a kind. But there are numerous hurdles in the way, reports Matilda Martin. For a seat a...

12 May

Time is running out for the UK to make a clear decision on where in the AI stack we can compete. Great to talk to @matildamartin.bsky.social for this piece on what comes next.

11 May

This shows just how limited our regulatory leverage is - US tech companies hold all the cards. But not everywhere. Some digital markets are genuinely still competitive. That's why the window to act is so narrow, we need to embed competition now, before AI follows the same path.

Hetan Shah@hetanshah.bsky.social

Scary diagram on how much of our daily lives would fall apart if somehow we were cut off from US tech. Tech sovereignty and resilience is going to be costly! www.ft.com/content/4c3a...

5 May

If AI starts to be a major source of information for people - it really matters where it gets information from. It is probably more complicated than AI having a clear political bias, but more transparency from AI companies is the bear minimum we need to understand what is going on here.

Institute for Public Policy Research

A part of this is due to where AI tools get their sources from. Earlier this year we published a report which found that the BBC and the Guardian dominated on AI platforms. Read more here: www.ippr.org/media-office...

Roa Powell reposted

Ethan Mollick

@emollick.bsky.social

5 May

Missing from the “will AI replace doctors?” debate is that doctors (and lawyers and psychologists and bankers) all vote & form the donor base to political parties & have deep community ties & sit in Congress. The government will largely determine what AI is allowed to do, no matter what it can do.

Roa Powell reposted

@ippr.org

28 Apr

Liz Kendall is right that our dependence on a handful of US tech companies is a national security risk. Three US tech companies control 80% of UK cloud infrastructure, we don’t have a single frontier AI company in the UK, and AI firms are increasingly deciding who gets access to their models. 🧵

POLITICO Europe@politico.eu

U.K. Technology Secretary Liz Kendall has vowed to work more closely with other “middle powers” to reduce the concentration of power in the AI market — arguing that doing so is in the interest of the U.S., too.

politico.euUK calls on ‘middle powers’ to counter over-concentration of AI controlThe technology secretary said Britain is making a “decisive move” towards backing its own startups.

lnkd.inLinkedInThis link will take you to a page that’s not on LinkedIn

20 Apr

In the US, the Democrats are already making AI a major part of their campaigns. I wrote about what Labour might learn from this, and why it isn't so easy for our politicians to ride any tidal wave of anti-AI sentiment. labourlist.org/2026/04/the-...

16 Apr

AI rose in salience more than any other issue for US voters last year, with backlash from creatives, X-risk activists, parents, workers, environmentalists and more We expect pressure to build in 2026 as AI gets even more capable, and @ippr.org we argue governments need to urgently spread the upside

🤖 | NEW REPORT: The UK risks a full-blown AI techlash unless the government shows who actually benefits. Our new report sets out a plan to make AI work for the public, not just big tech. Read here: www.ippr.org/articles/acc...

thetimes.comLabour ditches plan to let AI firms use copyrighted worksMinisters abandon ‘opt-out’ policy after backlash from Sir Elton John and Sir Paul McCartney over the ‘theft’ of creative material for software training

18 Mar

Great to see the UK ditch plans to give AI companies a free-for-all on how they use creators' data. But as government continues to sit on the fence, we are missing opportunities to shape a licensing market that is already emerging by supporting collective bargaining for creatives 1/2

18 Mar

As we documented at @ippr.org, the stand-off between creatives and AI companies is bad for everyone. It hurts AI accuracy (see examples of AI's patchy sourcing) and it hurts our news ecosystem (publishers predict a 43% fall in traffic in three years). 2/3

ippr.orgThe direction of AI innovation in the UK: Insights from a new database and a roadmap for reform | IPPRWith the latest models achieving top scores in scientific and diagnostic reasoning tests, they could usher in a new era of growth. In our previous report w

18 Mar

The best outcome would be collective licensing on fair terms - and a greater focus on AI outputs rather than just AI training - but at the moment AI companies hold all the power. 3/3

Roa Powell reposted

Carsten Jung

@carsjung.bsky.social

17 Mar

Replying to @carsjung.bsky.social

Particularly excited about the AI agenda taking some good shape. IPPR has long called for UK focussing on AI application layer, do more on sovereign AI, use procurement more actively + track AI economic impacts. Eg here: www.ippr.org/articles/the...

The BBC is facing a series of lose-lose choices. Partner with Big Tech and risk compromising what makes it unique. Or resist and fail to reach the audiences it exists to serve. @sofiahewson.bsky.social and I consider how Charter Renewal can help change this. 1/5

ippr.orgLevelling the playing field: The BBC, Big Tech, and the case for a bold charter | IPPRThe BBC faces a seemingly impossible choice. As trust in public institutions erodes and disinformation spreads online, its role to act in the public intere

📺 | NEW BLOG: How can the BBC compete in a media landscape increasingly dominated by global tech giants? Our latest analysis looks at what it will take to level the playing field - and why it matters for the future of UK public service media. www.ippr.org/articles/lev...

Charter renewal offers a chance to level the playing field with: 1. Conduct requirements for tech giants so the BBC can negotiate on fair terms 2. Prominence requirements so public service journalism is visible on social media 3. Funding and governance reform so the BBC can innovate on its own terms

The BBC shouldn't have to choose between reach and integrity, and a bold charter can help enable it to partner from a position of strength. Read the full blog from @ippr.org 5/5

Right now, the outcomes are bad whatever the BBC does: It can’t opt out – it blocks AI scraping but still has content misrepresented as chatbots turn to secondary sources And it struggles to partner effectively – its YouTube deal risks sending viewers to a commercial platform and weakening iPlayer

This is the predictable result of a massive power imbalance where tech platforms command vast resources and face minimal regulation, while the BBC operates on a shrinking budget and under intense scrutiny.

26 Feb

So far, AI companies have held all the power to decide how they use journalists' work. Today, the UK news industry have started to change this with a new coalition that gives journalism a collective voice to push back - exactly the sort of initiative we call for in our latest @ippr.org report. 1/3

26 Feb

Our research shows that AI can't currently be relied on as a good source of news - but that the answer isn't an endless standoff between publishers and tech companies. We need a fair deal that works for both sides and that means collective licensing of news content to be used in AI answers. 2/3

26 Feb

With the BBC, Guardian, Sky and more already signed up, this new coalition will help the news industry increase its bargaining power, but it won't be enough, and our report also maps the next steps government must take to ensure AI's impact on UK news is net positive 3/3

Roa Powell reposted

Carsten Jung

@carsjung.bsky.social

NEW: The policies of today will shape AI tools of the future. Let's make sure it improves - not destroys - our news environment. 👇

Roa Powell@roapowell.bsky.social

AI is transforming our information environment - but is it reliably delivering news content to the public? We tested four AI tools to find out and our results reveal worrying trends:

AI is transforming our information environment - but is it reliably delivering news content to the public? We tested four AI tools to find out and our results reveal worrying trends:

And, more work is needed to understand how commercial relationships between AI companies and news outlets impact prominence. Does having a licensing deal with OpenAI mean your content will be prioritised in AI answers? And will there be transparency around how these deals work?

We argue that a far better future is possible. To make it happen, we need: 1. AI companies to pay for news with government support for collective licensing so that a wide range of news outlets become visible. 2. Nutrition labels for AI answers and greater user choice on preferred sources.

AI is transforming our information environment - but is it reliably delivering news content to the public? We tested four AI tools to find out and our results reveal worrying trends:

AI also has the potential to create new winners and losers as news sites jostle for visibility. Generative engine optimisation (GEO) is starting to become the new SEO, but it won't always be the most trustworthy sites that win.

But - not all news sites want to be visible in AI answers. The BBC, the UK's most trusted news site, only shows up in half of AI tools. Murky rules on how content gets used mean that trying to block AI won't protect your news brand.

AI is transforming our information environment - but is it reliably delivering news content to the public? We tested four AI tools to find out and our results reveal worrying trends: