How lucky we are to live in interesting times. Over the past few weeks, we have witnessed the beginning of a historic transformation in world politics. 

Like everything else in global affairs, the art of the possible in AI policy is being redefined as the ground shifts beneath our feet. Stories matter, and the way we talk about things matters. Everyone wants the upcoming transformation into a world with AI to be “good”. But the definition and justifications of that “good” are changing.

We need to think carefully about paths forward, now that the political order is shifting and the very legitimacy of multilateral institutions is in question.

The old world is dying . . . 

Since the Bletchley Park summit in 2023, the international community has gathered around artificial intelligence from what we could call a “safetyist” viewpoint. The Bletchley Park and Seoul declarations highlighted the need for global cooperation and the importance of safety research; meanwhile, the International AI Safety Report was a global effort, uniting the best scientific understanding from around the world and pointing the way forward for risk-mitigating policy. 

All that multilateralism seems quaint and old-fashioned now, in the face of the USA’s astonishingly muscular reappearance on the world stage. 

In his first international speech, JD Vance dropped the mic on a full-on accelerationist stance for the USA (just before going on to urge Europe to embrace the far right in Munich). This was a shocking break for the former guardians of the free world. 

In a recent post, Anton Leicht analysed what this means for the future of AI policy. Two points from Anton’s post stood out to me particularly strongly:

• The notion of AI Safety has become “Dem-coded enough to suffer political retaliation, but not Dem-coded enough to be enacted on the coattails of general Democratic victories.”
• “The easiest trap to run into is reframing the movement without reconsidering the agenda. There’s been a growing consensus that ‘learning to speak Republican’ is advisable, and that maybe reframing AI safety, e.g. as AI security, could do the trick.”

I think this last one is a really important point. The policies AI safety folks used to advocate for – pauses, licensing, restrictions on who could develop AI, requirements for safety testing, MAGIC – all require adherence to multilateral institutions and mutlialteralist values to work. 

All those policies could be seen as “dem coded”. 

But I think that’s seeing things the wrong way. 

I think they’re more accurately described as “before-times-coded”. I am not alone in thinking that we’re seeing a shift away from the postwar multilateralist order, and not just from the USA. 

Vance may have shown the way forward with great audacity, but he is hardly treading the path alone. At the Paris AI Summit, Emmanuel Macron was absolutely not talking about slowing anything down. He announced €109 billion of funding for AI infrastructure to make France an “AI powerhouse” – part of the ‘third way’ for AI that Macron has been championing for years. 

Conversations I’ve had with some contacts in the Swedish government support this. “Don’t tell me this is a global problem requiring global solutions,” one told me, waving his hand in frustration. With Russia just across the Baltic, meddling daily in the political process, there’s no time to discuss slowing down. If you want a hearing in the halls of power, talk about unilateral policies that drive unilateral advantage, he said. 

So, where does this leave us when it comes to narratives around AI policy?

. . . And the new world struggles to be born.

This is our communications challenge. How do we express the concrete requirements of safe AI in a paradigm acceptable to the post-Paris system? We shouldn’t be clutching our pearls wishing for a return to the old world before 2025. We should be figuring out what we can do in the new one. And in the new world, despite everything we’ve seen in the last two months, I think there is a way for us to proceed. 

For example, AI Control is by definition a good thing. No one anywhere would prefer “out of control AI” to “AI we control”. This is as true in Washington as in Bejing – only the “we” is different. This is as true for business as it is for government. Banks, healthcare companies, and many other critical and regulated industries need AI they can trust to come to the right conclusions and not lead them into blind alleys or over cliff edges where they’ll hurt their customers – and perhaps get mired in lawsuits.

Protecting US AI companies from cyber threats is good for controllability, and consistent with the American POV.

Reskilling programmes, such as those proposed by the International AI Report, can still be acceptable, if they take on an economic nationalist tone. Not even authoritarian governments want large numbers of unemployed people – it is always and everywhere a recipe for unstable politics. (See: (1) COVID Pandemic → (2) Widespread joblessness → (3) Biden elected – an example that should be especially salient for US Republicans.)  

This last point really gives us an opportunity for content. Joblessness will become even more relevant, and quickly, as and if people start getting laid off through AI application in the US work force. Telling the stories of the displaced is somewhere that I would really like to explore. 

I agree that AI safety policy can’t go on like this. But it’s far from over.

Now is the time of monsters.

The rhetoric especially from the USA is essentially accelerationist, today. But as AI spreads through the economy and society, there will be pressure for that to change. That drives me to two conclusions (in bold below).

First, we need to be ready to tell and spread the right stories, as those risks become more salient and instantiate into damage.

By analogy, climate change remained a difficult, systemic abstraction until people in the rich world woke up to find their houses under water, or burning in a wildfire. Now that more obvious and damaging consequences of the global temperature rise are obvious, political action is based on a swell of public grievance. It is difficult to photograph a global temperature rise of 1.5 degrees, but seeing a video of someone crying as their house burns tends to concentrate the mind, and creates emotional impact that can drive action.

Grievance is a powerful emotion, and it’s increasingly common. Recent research shows us that almost two thirds of people worldwide feel this way (see slide 17 of this report). That means they feel they feel wronged by the establishment. We can work with that.   

The accelerationist anti-establishment approach we’re seeing from Elon Musk in the USA might actually make this easier. There will likely be some thousands of jobless former federal workers by the end of the year. Among their stories will be some that we can tell and employ to build a movement. This is how we build public pressure for change, story by story, brick by brick.

That’s why, as the consequences of misaligned AI start appearing, I think we will need to be at the ready to tell those stories with impact. 

My second conclusion is that we will still need to position our safety discourse within the dominant political paradigm, or it won’t go as far as it could.

Even faced with significant consequences from AI, there are limits to how far the Trump administration can change their rhetoric. They can’t compromise their core values, or they lose all legitimacy. Eschewing international institutions in favour of a unilateral focus on America First is one of those core values.

And remember that the political shift we’re seeing isn’t an American one. It’s global, as we saw in Germany a few weeks ago. Even NATO is in doubt – the man who used to run it says he is “struggling to comprehend a transatlantic relationship that is crumbling before our eyes.”

This is a historical, political shift in balance that we must take into account. Previous AI Safety planning included things like an “IAEA for AI” under the UN, or the even more restrictive MAGIC proposal. I don’t think it’s realistic to push for those things now – or it has, at the very least, become much much harder. Perhaps we will see something like MAGIC – but only if it is a Manhattan-project-like initiative at a secure location in the USA, aimed at creating controllable superintelligence specifically before China does.

My point here is that we need to be smart about the way we speak truth to power. It needs to be a truth they can fit into their narrative of power, or it’ll get rejected and we miss an opportunity for impact.

So – I think the time of AI Safety may be passing. I think the time of Beneficial AI is beginning.

Many thanks to Rowan Emslie and Anton Leicht, whose posts were inspirational to this.

TL;DR: Major AI companies are racing to build the world’s biggest data centers in a hail mary attempt at AGI. We don’t know whether that will be enough to build AGI. But what seems certain, if you look 2 moves ahead, is that we’re already in the first stages of an AI-driven cold war with China. That story is going to be the backdrop of the next decade. This post is based on a talk I gave at the Berghs School of Communication Unconference on AI.

There has been an explosion of news about this over the last week. AI expert Gary Marcus claims this is the end for growing AGI through LLMs. Just yesterday, Reuters broke this story about how the big AI labs are having to change research track. Online communities are buzzing with back and forth debate on this. As usual online, there’s a lot of hype and noise out there.

But it’s really important for us to understand this story, as communicators. It’s not just an academic question.

I’ve written before about how the debate about AI is going to affect the operating environment for all of us communicators, just as sustainability has. What happens in AI is going to affect every aspect of the information environment we work in.

So even if you don’t work in AI policy, this is important to know.

In this post, we will:

• Define what AGI is, and AI winter would be
• Look at the promises and hype around AGI
• Lay out what would be necessary to make AGI happen
• Examine the challenges we know stand in the way of AGI
• See how this leads almost inevitably to great power competition between China and the USA

Are we on the verge of AGI?

Building Artificial General Intelligence, AGI, is the explicit goal of the major AI companies. This means general-purpose AI with human-like adaptability and reasoning; a machine intelligence advanced enough that it can replace any remote worker on a 1:1 basis.

If we succeeded in building AIs like this, they could be run in thousands or millions of instances, round the clock, creating what Dario Amodei has called “A country of geniuses in a datacenter”.

He and other AI leaders are bullish about what such a concentration of geniuses would enable. In fact, the future is so bright it’s blinding. Huge economic growth, medical breakthroughs up to and including curing death, merging with the AI, living forever in simulated bliss and colonizing the reachable universe. The future AGI would bring about is so wild that the human mind recoils. This is a bug in human reasoning – we’re famously bad at appreciating exponential growth.

It’s easier to dismiss as hype.

And it seems many people are reacting to the wild visions of the AGI-powered future as hype.

The Gartner Hype Cycle for 2024 has GenAI sliding down the trough of disillusionment and AGI not far off the peak.

Figure SOURCE

. . . So it’s time for an AI Winter?

Each previous AI winter followed a period of hype, often driven by technological limits and unrealistic promises.

In the history of AI development, we’ve had a few of these periods of stagnation in AI development, marked by waning enthusiasm and investment due to unmet expectations.

Figure SOURCE

But is it different this time? What’s the case for AGI soon?

In a nutshell: scaling laws.

“To a shocking degree of precision, the more compute and data available, the better [AI] gets at helping people solve hard problems.” Sam Altman

Scaling Laws FTW

What we’ve discovered, and what Altman was talking about in that quote above, is that the more compute you dedicate to training AI, and the more data you train it on, the better it gets.

With incredibly consistent results.

See this paper for a technical discussion of this.

Figure SOURCE

These graphs show how test loss decreases smoothly and predictably as you scale up inputs to creating AI models. Note though, that the x-axis on those graphs isn’t linear. It’s exponential. Each step to the right involves a 10x increase in compute, dataset, or parameters.

So the scaling laws appear to hold, but you have to massively scale up your inputs to make them do so.

The real question is – how far do they hold? And will they hold long enough – is this all we need to create AGI?

Previously, Dario Amodei, Sam Altman and other AI leaders have appeared quite optimistic on this in public.

But they would.

Their investments depend on this enthusiasm. Nobody is going to give you a hundred billion dollars if you aren’t 100% committed to your crazy idea.

That is a Dr. Evil amount of money.

The challenges on the road to AGI

Scaling laws have held so far. And if they continue to hold – and that’s a big ‘if’ – then all we need to do is train AI on more and more compute, and it’ll get smarter and smarter.

Well – ‘all we need to do’ is glossing over a lot. Even if it’s just a question of training scale, there are some challenges ahead.

Let’s take a look at them.

Chips are hard to make . . .

Making the chips necessary for advanced AI training is technically very difficult.

They are among the more sophisticated objects that humans have ever built, and only a handful of global players are actually capable of developing them, creating a bottleneck that affects the entire industry. (This concentration makes chips a key point of leverage in AI governance. Unlike data or software, chips are tangible and restricted by physical limits in manufacturing and distribution.)

The upshot of this concentration means that if you’re trying to 10x the compute in your next AI training run, you can’t just go down to Walmart and stock up on the chips you need.

Outside the industry, it’s easy to miss just how hard it is, and how expensive it is, to get enough chips for scaling AI models. These frontier devices are so advanced that physically delivering the first of a new generation of chips is considered worthy of a multi-CEO photo op.

So if you’re an AI developer, and you’re trying to get enough chips to increase your training runs 10x or 100x, this isn’t a trivial expansion—it’s a multibillion-dollar infrastructure investment that only a few global players can afford.

. . . Data centers are power-hungry . . .

Chips go in data centers, and as demand for data centers is exploding, their power requirements are soaring. This is the second bottleneck to AGI development.

Currently, U.S. data centers consume about 4% of national energy production, but Goldman Sachs forecasts that this figure could double to 8% by 2030.

Some experts, like Leopold Aschenbrenner, predict it could reach as high as 20%, requiring an additional 100 GW of energy capacity, at a cost of over a trillion dollars (larger than the budget for the US military).

If you think this is outrageous – well, it is. But it’s already happening.

In Ireland, for example, data centers consumed 18% of the nation’s electricity in 2022, matching the energy use of all urban housing combined.

That’s right, data centers in Ireland consume as much energy as all the houses in Irish cities. Not in the future – today.

This strain has led EirGrid, Ireland’s state power provider, to impose a moratorium on new data centers in the Greater Dublin Area until at least 2028.

. . . and all this is going to cost a lot of money.

Accommodating these demands will require vast levels of investment. OpenAI and Microsoft are already planning the Stargate data center, at an eye-watering cost of USD $100 billion.

This is a taste of things to come.

The level of investment we’re looking at here means the upcoming infrastructure push for data centers and compute could surpass even iconic projects like the Apollo Program, the Manhattan Project, or Britain’s railway expansion in the 19th century.

For historical context, Britain’s railway expansion from 1841 to 1850 consumed roughly 40% of British GDP, a scale that would equal around $11 trillion for the U.S. over a decade.

The Apollo Program and Manhattan Project—two of the most ambitious American undertakings—required funding equivalent to around 3% and 2% of GDP, respectively.

It is highly relevant that both the Manhattan Project and the Apollo Programme were propelled by the USA racing against an external rival – the Nazis, and the Soviets, respectively.

It seems like this pattern is about to repeat itself, between the USA and China.

But there’s one more question we need to answer before we talk about the coming Cold War.

What if we need new science, too?

Can we create AGI if we ‘just’ scale up our existing systems? Large Language Models, our current generation of AI, are built on several breakthroughs in computer science, such as

• Deep neural networks: machine learning
• Convolutional neural networks: computer vision
• Transformers: language modelling

Do we need another breakthrough on the level of these to get to AGI? Or maybe several?

The simple answer to this is – we don’t know.

Leading AI experts offer a mixed view:

[Demis Hassabis, founder of DeepMind] “I think there’s a lot of uncertainty over how many more breakthroughs are required to get to AGI, big, big breakthroughs — innovative breakthroughs — versus just scaling up existing solutions. And I think it very much depends on that in terms of timeframe. Obviously, if there are a lot of breakthroughs still required, those are a lot harder to do and take a lot longer. But right now, I would not be surprised if we approached something like AGI or AGI-like in the next decade.”

[Andrew Ng, Founder of Google Brain] “I see no reason why we won’t get there someday. But someday feels quite far, and I’m very confident that, if the only recipe is scaling up existing transformer networks, I don’t think that will get us there. We still need additional technical breakthroughs.”

[Yann LeCun, Head of AI at Meta] “It’s not as if we’re going to be able to scale them up and train them with more data, with bigger computers, and reach human intelligence. This is not going to happen. What’s going to happen is that we’re going to have to discover new technology, new architectures of those systems. Once we figure out how to build machines so they can understand the world — remember, plan and reason — then we’ll have a path towards human-level intelligence.”

Many other AI researchers and thought leaders (including Sam Altman) have gone on the record sying that new science will be needed. But there are already many promising avenues of research, and some of the smartest computer scientists in the world are working on this.

We don’t know when it will happen. But here’s what we do know.

Here is a breakdown of what is extremely likely:

Politically, with Trump’s win in the USA, the environment for leading-edge AI development in the USA is going to change. Shakeel Hashim at Transformer had an excellent breakdown of this. Trump has said he’ll rescind the Biden Executive Order on AI, which is currently the US’ leading document on AI policy. Elon Musk will likely be at Trump’s elbow as de facto AI advisor – Musk is pro-AI safety, but he’s also pro-business and definitely pro USA, so it’s unlikely that this will result in a slowdown of AI development in the USA.

In business, the above makes it likely that we’re about to see something like the unprecedented investments in compute and energy that I outlined above.

Internationally, this means that we’re about to enter a new cold war.

Great power competition

Under Trump’s presidency the pace of US investment in AI technology is going to increase, not slow down. There’s no way that China is sitting this one out. For either superpower, letting the other develop AGI without competing is unthinkable.

Think of what a military technology advantage can do. In 1991, the American-led coalition in the first Gulf war obliterated the Iraqi military in 100 hours of ground operations, suffering 292 dead, compared to 20-50 thousand Iraqi dead. The coalition lost 31 tanks, the Iraqis 3000. That is what a technological advantage can give you in a shooting war.

This is to say nothing of the potential advantages that await those who can successfully deploy AI in cyberwarfare or online espionage. This type of mathematics tends to concentrate the mind.

China has mobilized substantial resources towards AI research, in defiance of the USA’s embargo on exporting advanced chips there. They’re also pushing hard to get to AGI first.

This has obvious echoes of the nuclear arms race from the first cold war. And, just as in that war, you can bet that both the USA and Chinese intelligence and diplomatic corps are focusing on this problem too. Technology can be stolen. Imports can be blocked.

And, in case intelligence and diplomacy fail, powerplants can be sabotaged. Mines producing rare metals can be occupied. Shipping lanes can be blockaded.

Datacentres can be bombed.

This is one reason why TSMC, the world’s largest producer of computer chips, is building facilities in the continental USA.

The overwhelming advantage accruing to the country that first develops AGI motivates the kind of existential threat that led to doctrines from the first cold war like mutually assured destruction.

We’ve been in this dynamic before.

Looks like we’re going to be in it again.

Effects on the media environment

Separately to the probability for a new cold war, we should be worried about how information technology has affected democracies and autocracies.

In the 1990’s, we thought that the spreading internet would inevitably open up China and Russia.

Instead, we’ve seen how China has applied information technology to build a Great Firewall and create an online climate structured to the values of the Party.

Meanwhile, far from strengthening our own democracies, we’ve seen social media intensify political divisions and make stable government more difficult – aided, in fact, by the intelligence services of Russia and other authoritarian actors throwing their disinformation bombs into the factional information mix.

Why would we think that AI will reverse this trend?

Thirty years ago, Francis Fukuyama wrote about The End of History and the final triumph of liberal democracy. That hasn’t happened yet. In fact, since Fukuyama wrote, the adoption of information technology has weakened democracies and strengthened authoritarian regimes. With the probability of AI-supported disinformation in the mix, this situation is only going to become more volatile.

Conclusion

We don’t know for sure what’s coming next when it comes to AI. We could see AGI as soon as next year, if the most optimistic predictions come true. Alternatively, it might be so hard to do that most of us won’t see it in our lifetimes. Most AI researchers seem to think it’s quite likely we’ll have AGI at some point in the next decade, maybe two.

Whatever these timelines turn out to be, we’re in for a wild ride. Progress won’t slow down, despite what some are asking for. No – rather the opposite. Change will never be this slow again.

For us as communicators, it means we need to stay on this issue and get fluent about it, because it’s going to form the backdrop of the global political and business environment we all operate in.

On December 6th I spoke at the IAB’s “8 Minutes” seminar in Stockholm. This video of my talk is a high-speed overview of the three arguments that I believe will shape the policy and communications environment for all of us in the near future.

If the whirlwind of 2023 is anything to go by, 2024 will see an even louder argument about how we should adapt to AI around the world. And this is especially true as we’re about to enter the biggest election year in history, with the US, the EU, and India among others all going to the polls.

Understanding the basics of AI policy is going to be important not just for tech geeks, but for all of us trying to make a living in an information environment that we’ll have to share with the Balenciaga Pope.

Sustainability for bits, not atoms

Think about how sustainability has transformed business and communications. From a fringe concern, the discussion about sustainability has rightfully become central to business strategy. Now the conversation about climate is commonplace in boardrooms. Carbon emissions are related to corporate risks, and the road to net-zero is a question of strategic necessity. 

AI safety will soon go through a similar cultural transformation. As a general-purpose technology, AI is likely to affect our entire economy. Few organisations, if any, will be able to continue functioning while ignoring its effects. 

In the words of Marc Andreesen, AI is about to become “the control layer for the world.” Yuval Noah Harari describes AI as affecting “the operating system of our civilisation.” Adapting such power means accepting the attendant responsibility. Even something as mundane as an AI app that suggests recipes can be dangerous. (In this example, by suggesting recipes that would actually have created chemical weapons . . . !) 

That is why business leaders will soon have to focus on AI governance: it is an equivalent of sustainability for bits, not atoms. A technology this powerful is too dangerous to let develop without some thought and direction. It is time for business leaders everywhere to get serious about how we will govern this technology, both within our organisations and in society at large. 

The false division and the risks of polarisation

To understand this coming cultural transformation, we must understand the dynamics of the discussion around AI.

There are three main perspectives on AI, each with its own values and visions of risk. These are the existential risk perspective, prioritising safety; the AI ethics perspective, prioritising justice; and the accelerationist perspective, prioritising prosperity. 

Often these are presented as rival factions. In the media, AI Doomers seem to dismiss woke social justice commentators; meanwhile both are derided by the venture capitalists desperate to beat China in the race to build superintelligence. 

Dramatic as it is, this is a false trichotomy. Each perspective is valuable. 

Business leaders would do well to learn from all three in the effort to arrive at sensible AI policy.

Existential Risk: AI as a Threat to Humanity

Could AI actually destroy civilization and end the world? 

We can’t rule it out.  

Humans have destroyed many other species, not because we are malevolent, but because we are more intelligent. If we build machines smarter than we are, we had better be sure that they are well governed. The risks of accident or misuse will grow as these machines take on critical roles managing our infrastructure, financial markets, and information systems. It is even possible, in principle, that superintelligent machines, with goals we can barely understand, could destroy us by accident.

There is disagreement here about the likelihood of the risks. But with the risks of widespread damage on the line, this isn’t something that we can simply ignore. Many leading thinkers in the field have come out to support this view. We all have a responsibility to promote safe AI. 

Justice and Social Impact: Mitigating Harm and Inequality

Another perspective focuses on the immediate social harms caused by AI, such as biased decision-making, perpetuating prejudice, and amplifying inequality. 

There are manifold examples of AI systems already causing this type of harm today. AI’s impact on democracy, manipulation through social media, and the creation of deepfakes raise concerns about social disruption and the need to address injustice. 

Implementing AI systems in your organisation means taking the risks of bias and distortion seriously. It means accepting a responsibility to reduce harms wherever we can find them. 

Prosperity and Rapid Innovation: Balancing Benefits and Risks

These two points bring us to a delicate balance. Each highlights real risks. But we must balance them with the prospect of realising the value of innovation. There is another risk, from regulation itself, since it might slow down innovation. 

From this viewpoint the stakes are also very high. Since AI advances tend to build on each other, there is a tremendous potential first-mover advantage. At least one AI lab seems to agree: a leaked pitch deck from Anthropic claims that “companies that train the best 2025/26 models will be too far ahead for anyone to catch up”.

With this in mind, there are reasons that slowing down progress is its own risk. Personally, I believe in liberal, democratic values (as, I imagine, do most of you reading this). But whoever trains the most powerful AI in the next few years could end up having a disproportionate effect on world culture for a very long time. Perhaps centuries. What if the future is built upon AI that is trained to censor public debate and hold up the primacy of a political party? That also seems a risk worth mitigating against.

So what are the lessons for business leaders?

So, is AI a future existential risk? Or a present social justice one? Or is it a huge opportunity that we need to rush to apply?

The answer is . . . yes. It’s all three. And adopting responsible AI policy is the responsibility of every leader, just as it has become the responsibility of every leader to adopt effective sustainability policy for their organisation. 

Sustainability policy for atoms, AI policy for bits. 

The time to start getting involved in crafting AI policy is now. International momentum is gathering for regulation, national legislation will soon appear. More importantly, all of this is part of a sweeping cultural change that is happening incredibly fast. 

Artificial intelligence will undoubtedly transform our lives and work over the course of coming decades. We all have a role to play in ensuring it is a transformation that benefits everyone with safe and just AI that drives prosperity. 

This is a condensed version of the full white paper on the politics of AI governance available here.

Yesterday I joined the Unconference on AI at Berghs School of Communication, one of the world’s leading creative industry schools. It was a fascinating chance to hear from professors, students and practitioners trying to figure out how to incorporate AI tools in their work. The pressure is on; according to recent research, it seems likely the communications and marketing industry will be more affected than most.

Some patterns are starting to emerge.

We all (almost) feel augmented.

I’ve become fond of saying “With ChatGPT, I feel like I have an intern now.” Turns out, the thought is quite common: I heard it from speaker after speaker.

Professional communicators especially feel like they are augmented, as they can use their experience to frame specific tasks for a generative AI assistant.

The importance of experience and the value of art direction are important here. In an era where content production itself can be automated, much more depends on having the ability to brief effectively, and on having the taste and discernment to review and refine work. We call this art direction and editing, and its relative value just increased substantially.

But we don’t all feel augmented.

What surprised me was speaking with some teachers at Berghs and learning that they notice that some of their students are reluctant to engage. Some are diving in, but some say it feels like cheating, and they prefer ‘pure’ creativity instead.

Standard workflows are developing.

 Meanwhile, in the industry, it’s full steam ahead. Many speakers described a similar intermodal workflow for generating campaigns:

1. Using Chat GPT for client industry research to identify key topics
2. Using Chat GPT to ideate on those topics for insights, base slogans etc. – essentially an accelerated brainstorming/ideation process
3. Working up this material into core propositions “manually”
4. Developing copy with Chat GPT
5. Using that copy as a base, asking ChatGPT to create prompts for Midjourney to generate image creative
6. Integrating the copy and image into campaign materials

This workflow can generate relatively high-quality marketing materials for standard text-and-image campaigns extremely quickly.

One speaker told us that he had been able to generate a whole client campaign in half an hour – something that would have taken three days earlier, and required hiring an illustrator or graphic designer.

We’re going to have to change our business models.

So, you can generate three days’ work in half an hour. Maybe your time savings won’t be quite so radical, maybe you can do three days’ work in a day. Regardless, I think we’re all starting to realize that we can work much faster, if we’re doing the same type of work we did before.

For an industry where the standard is that we are paid by the hour, that poses some pretty big questions.

• Does your client approve the work at 09:30 and expect a full campaign delivered by lunchtime?
• Do you just deliver the same campaign you could have done without AI, price it at 3 days of your time, and go to the pub?
• Do you put three days of work into it anyhow, expanding the campaign and delivering more in the time than ever before possible?
• If so, what additional features and deliverables are you adding to the campaign to make it worth three days of AI-augmented work?

So, this is the real question here. How do we deliver new types of communication work that weren’t possible before?

Not just faster. Different. 

When we started delivering social media campaigns, over a decade ago, we discovered that the technology enabled new creative forms of communication. One of these was influence work and influencer marketing, an emergent mode of marketing communications native to the web 2.0 era. We expanded the old repertoire of copy, visual and video with new ways of reaching people and engaging them.

We haven’t worked out yet what the equivalent of this will be in the age of generative AI. 

We clearly have our work cut out for us.

Social media and search have transformed the way information spreads through society, and not always for the better. 

We’ve all been talking about community fragmentation due to algorithmic media for a long time. 

Today we live with headlines about political fragmentation and polarisation, the rise of conspiracy theories, and disinformation campaigns that strive to capitalise on these fault lines. 

Every news source bursts with data, each push notification in our pockets adding a vibrating note to the ambiance of threat. 

People are increasingly worried about losing their freedoms as citizens (65%), climate change (75%) and most of all job losses (85%).

When faced with such threats to lives and livelihoods, fear is a natural reaction. And fear tends to move decision-making down the Maslowian hierarchy of needs. The lower down that hierarchy you’re operating, the more likely you are to tend towards more emotional, less rational decision-making.

With all these informational threats on our radar, it’s not surprising that social trust has collapsed. Research shows that as a result of our challenging information environment, distrust has become the default setting in society. 

Rally Round the Family

As a result of this pressure, people are trusting distant authority sources less, and the people close to them more. The mechanics of online platforms exacerbate this trend, showing us more content from people we are likely to agree with. 

The human reaction is understandable. In uncertain times, we connect more with our communities. We spend more time and attention on like-minded people united by shared interests, opinions and values, and we’re less likely to trust those who don’t share our interests, opinions and values. 

The importance of trust can’t be overstated here. We choose the people (and therefore the communities) we engage in, and that defines our media bubbles. 

In other words, who we trust defines the media bubbles we live in. 

The upshot of that is that our media bubbles are actually created and defined by influence.

How so? 

The Role of Influence

Online communities don’t create content by themselves. They’re platforms. Users create the content.

The activity of any online community is driven by its members. 

And we know that not all members of an online community tend to create content at equal volume. Some of those members are more active, more central, than others. Those are the influencers. They capture attention, they become rallying points for followers – and these followers become a community. In a sense, influencers become the focal points for online communities, and online communities are built up, and perpetuated, through the actions of the influencers at their core.

So here’s what we’re seeing: 

1. Social trust is collapsing, people are worried about multiple perceived threats 
2. Attention and time is naturally focusing on people who are close to us – in attitude and values if not in physical reality 
3. This creates media bubbles, a fragmented media environment 
4. But what actually happens in those media bubbles is defined by the users – i.e. content is created disproportionately by a few prolific creators at the core. 

Media bubbles are, in a very real sense, defined by the influencers at their core. 

In other words, influencers and media bubbles are linked as central features of an information environment dominated by algorithmic media. They are actually inseparable; you can’t have one without the other. 

This is why understanding influence is critical to penetrating media bubbles and earning trust. 

The Promise of Influence

This is the promise of influence. In a difficult information environment, mapping and understanding communities of interest isn’t just useful. 

It’s essential.

As communicators we must accurately map the communities we need to mobilise. We must find, understand, and work with people the community is focused on. In a world of rising distrust and hardening media bubbles, we can only ensure our message is trusted with a data-driven understanding of influence.

It is time for a code of ethics for the use of persuasive technology in marketing and communications.

Where does effective marketing end, and unacceptable manipulation begin?

[Photo by Sagar Dani on Unsplash]

This is a question I’ve been asking myself often lately. I make my living trying to affect the choices people make. Everyone in the marketing and communications industry does. Clients hire us to change the way people choose, act, talk, and think. Often that’s a question of choosing one brand of product over another, but the range of tasks we take on is vast, and we use some pretty advanced technology to do this work.

Concerns about the persuasive power of technology aren’t new. Berdichevsky and Neuenschwander coined the term “persuasive technology” over 20 years ago. But there are a few things we can do now that we couldn’t do back in 1999.

Let’s have a look at some persuasive technology I have access to, as an ordinary member of the marketing and communications industry.

• Mass surveillance: the ability to monitor the activity and interactions of large numbers of people though aggregated social media and mobile data
• Analytics/predictive technology: the ability to segment large audiences into finely defined persuadable groups, through big-data analytical techniques
• Behavioural segmentation and optimization: the ability to fine-tune messaging to persuade specific types or groups of people through the application of behavioural science techniques such as personality mapping
• Micro-targeting: the ability to make messages reach certain types of people at a critical point in time, such as close to a purchase decision, or in a particular emotional state

None of these techniques is intrinsically good or bad. And for our clients, they represent an incredibly valuable set of capabilities. Take them together and we can do some really, really sophisticated campaigns aimed at valuable outcomes. When we deploy these techniques appropriately, we can connect people with opportunities that improve their lives and the lives of those around them, while growing economies. There’s no question that we should stop using them.

But how do we know when we are taking it too far?

When I tell colleagues and clients about everything that we can do, I often get the same question:

Isn’t this all very Cambridge Analytica?

[Photo by Craig Whitehead on Unsplash]

It’s a valid question. Cambridge Analytica famously used advanced persuasive technologies to shift elections, including the Brexit referendum in the UK and Trump’s election in the USA.

So what makes us any different? If Cambridge Analytica had been fighting for the liberal internationalist side – the Remain Campaign or Clinton’s presidential campaign – would what they did have been acceptable? Would Democrats and Remainers be cheering them as liberal heroes?

I certainly hope not. It’s no secret that I’m a liberal internationalist, but even if Cambridge Analytica had been pulling for the side I agreed with, they still did serious wrong. First, they built their segmentation and targeting campaigns on stolen data. Second, they lied in the messaging that they distributed.

And that shows us something else. Regardless of the aim of a given persuasive campaign, there is an ethical way to use persuasive technology – and an unethical way.

It’s time for a code of ethics for the use of persuasive technology in the Marketing and Communications industry.

[Photo by Rob Curran on Unsplash]

In the industry and in society at large, more and more attention is being focused on the ethical use of the technology of persuasion. Trust in institutions including the media and large tech companies is eroding.

There are already regulations and codes of ethics for advertising, for PR and for AI in general. However, there is no specific code or principles for the technology of persuasion, or for the use of AI in marketing and communications.

This is why I believe it is time for a code of ethics for the use of AI in marketing. National or international legislation is far behind the curve of change for this topic. It is time for the industry to come together and agree on a set of principles that define where effective marketing ends and unacceptable manipulation begins.

I’m not here to lay this code out alone – that is far beyond my authority or my expertise. This will take a concerted effort from thought leaders in the marketing and communications industry, but also beyond it: in academia, in ethics, in tech companies, in governments and the military.

Nevertheless, this is something that must be done, and it must be done now. Our capabilities and our power are growing day by day – I see the truth of this at work every day.

Nothing can be more central to who we are than the choices we make. Those choices have always been shaped by what we read, see and hear. But until recently what we all read, see and hear couldn’t be fine-tuned with such power and detail. It’s time to take a look at that power and decide how we should – and should not – use it.

It’s like my colleague Penny Lipsham said: “You can take the political ads out of Twitter, but you can’t take the politics out of Twitter.”

You can ban political ads – if you define them narrowly, as ads paid for by a registered political party.

But what about issue ads?

What about Greenpeace urging people to take action against climate change? Some political parties have climate issues at the heart of their electoral platforms. Does that mean Greenpeace is spreading a political message?

And then there’s business and social purpose.

What about Nike and Colin Kaepernic? That campaign set the company squarely on the Democrat side of the political divide in the USA. Republicans across the country were burning their Nikes after that moment. Nike had correctly calculated that the sales boost from aligning with an urban, liberal audience outweighed the lost custom from Republican-leaning buyers. But you’d have to squint pretty hard to not see that as a political move.

Or what about HSBC’s recent We Are Not an Island campaign?

They didn’t come straight out and say “Bollocks to Brexit” but if you look at the comments under that video, you’ll see the evidence: people see this as the company taking a stand on Brexit. Like it or not, this is the cognitive landscape they are playing in. [Disclosure: HSBC is an Edelman client.]

This is only going to grow.

We’re seeing a growth in purpose-driven marketing worldwide, because we’re seeing a cultural shift towards companies focusing on more than just the bottom line. Our own Trust Barometer research shows that people expect this of companies (especially their employers):

• A major consideration for brand purchase is now “I must be able to trust the brand to do what is right,” at 81 percent. 
• More than 70 percent link purchase to considerations that historically were tied to trust in corporations, including supply chain, reputation, values, environmental impact and customer before profit. 
• 53 percent of consumers agree that every brand has a responsibility to get involved in at least one social issue that does not directly impact its business.  

And this summer, the Business Roundtable of 181 CEOs agreed with that ethos. So we’re likely to see more and more businesses get involved in social issues – thus tying marketing and communications closer and closer to issues that matter in civil society, and inevitably, to politics.

That’s why banning political ads on Twitter won’t take politics out of Twitter.

It seems like it’s the wrong answer to the issue. It’s a fact that Twitter and Facebook have become critical arenas of public discourse. Politics and civil society conversations are going to keep happening there.

The issue isn’t that these conversations are happening there – the issue is that they should be happening there transparently, honestly, and truthfully.

I’d say political ads aren’t a problem in and of themselves. But ads based on lies are a problem. Manipulative ads that exploit AI-driven behavioural microtargeting are a problem. Opaque systems that allow foreign manipulation of all the above weaknesses are definitely a problem.

Banning all political ads just doesn’t seem to be addressing the problem effectively. 

Influence is complex. But in the communication/marketing industry, many people still seem simply to equate reach with influence. In this logic, if someone has two million followers, then they must be twice as influential as someone with a million followers. Influencer pay is certainly a function of overall follower count, a correlation that has been justly criticized.

But we all know there’s more to influence than reach.

There are three other things, actually. Together, they are the four dimensions of influence.

They are Relevance, Authority, Reach and Accessibility.

Together, these dimensions form a complete model of social influence, both online and offline.

When we are doing influencer selection work, we consider each dimension and then bring them together to form a composite assessment of an influencer’s appropriateness for a particular brand or campaign.

Relevance

Do this person’s values overlap with the brand’s? Are they on-topic?

Is this person already speaking, publishing, posting and sharing material that is naturally relevant to the brand? To the target audience? Is it on-topic for the current activation or campaign? If you were to put the brand’s logo on the end-frame of one of this influencer’s videos or posts, would it look natural?

What about their personality? Does it overlap with the brand’s identity and the likely audience profile? With psychometric profiling technology we can get very specific and objective here.

These are all factors that will affect the person’s relevance score.

Authority

How trusted is this person’s expertise in this subject area?

Are they discussed or linked to by peers or media as someone who knows about the topic at hand? Do they have professional experience in the field? Are they perceived as an expert? These are some of the factors will increase their Authority score.

Reach

How many members of the target audience are paying attention to this person?

Not just total followers – that’s just a lifetime number of how many people have clicked ‘subscribe’ at some point. We have the technology to get more specific. How many people are watching/reading/listening right now? Are they the kind of people we need to reach?With the right tools for demographic analysis we can get very specific on this.

In many ways this is the least important dimension of the four, because reach can be bought very effectively through targeted paid spend.

Accessibility

How willing is this person to work with the brand?

This is a function of the influencer’s position and of their attitude towards the brand. They may be very popular and booked months in advance. They may command high fees. They may be a total fanboy/girl of a rival brand. They may be employed in a position that restricts their ability to speak freely. Any one of these factors would reduce their Accessibility score.

The RARA system: Quantified Influence

Together, we call this the RARA system. It means we can be very specific and very targeted with clients when we recommend influencers to work with.

The major advantage here – and the opportunity for value to the client – is that we can actually measure each dimension of influence, based on data relating to each one, processed through proprietary algorithms. The following inputs give us a score from 0 to 100 for any influencer for each of the four dimensions:

• Relevance can be measured through keyword matching, entity-defining KPIs, and many other sources.
• Authority can be measured through contextual search ranking, social media engagement statistics, and so on.
• Reach is well-studied, and we have the technology to be quite specific about exactly what type of person is paying attention to which influencer and when.
• Finally, Accessibility can also be measured as a function of influencer location, existing brand partnerships and fee expectations.

The specific details of our method are confidential, but the system is robust and works across languages, regions, and can account for influence both online and off.

This means influence can be effectively quantified. And if it can be quantified, it can be mapped, like this:

Don’t get too hung up on the names (they’re from randomnamegenerator.com). The point here is to illustrate what a landscape of influence looks like, around a given topic or brand area. These could be pet bloggers, video game streamers, or politicians and journalists. The system works equally well for all of them.

The horizontal axis measures brand relevance. The further to the right, the more this person is creating content that fits well with the goals of the specific client communications campaign.

The vertical axis  measures topic authority. The higher up, the more this person is trusted to speak on the subject of the communications challenge.

The size of each bubble measures reach. Larger bubbles indicate this person has a larger engaged audience.

Finally the hue of the bubble measures accessibility. The darker the bubble, the more that person is willing to work with the brand and shares brand integrity.

Contextual Influence

This system is contextual. Any individual’s scores will vary depending on the specific topic and activation their scores are being mapped against.

• Relevance varies according to brand and communications topic. An influencer could be on-brand for one brand but irrelevant for another.
• Authority varies according to topic. I might trust Zoella to recommend a makeup brand, but not a motor oil.
• Reach varies according to target audience. An influencer with 10 million teenage girls following her has a very low reach if the brand partner is an IT security firm, for example.
• Accessibilty varies according to brand partnership. An influencer might be very open to working with a techn brand but now a mining company, for instance.

So while it’s not possible to give an individual general scores, it is possible to assess any individual’s appropriateness very precisely for any specific brand partnership.

Therefore, this method offers a robust and conclusive method of determining which individuals should be partnered with for a given communications objective.

Conclusion

The RARA system solves the problem of deciding which influencer a brand should partner with.

By breaking influence down into four dimensions, we isolate the various components of influence and make it easier to assess.

Furthermore, through our data-driven approach and proprietary algorithms,  we can effectively and consistently quantify influence. This makes it possible to objectively compare influencers from different regions or platforms on a like-for-like basis and make appropriate partnering and budgeting decisions.

If you’d like to find out more about the RARA system and how we apply it to client work at Edelman, let me know.

Over the past few months, Facebook has made three changes to the way influencer marketing works. Together, these changes give Facebook Inc. a very strong position in influencer marketing, and a good shot at eating a big slice of the global influencer marketing money pie.

So what have they done?

• First, they’ve made it impossible to perform influencer marketing on Facebook and Instagram without notifying Facebook itself (Branded Partner Feature).
• Second, they’ve made it as easy as possible for creators to make loads and loads of high-quality content (Facebook Creator App).
• Third, they’ve opened a function to allow brands to easily find and engage content creators in any niche (Branded Content Matching).

This is a brutal 3-hit combo. With these changes Facebook has laid the groundwork to take a huge cut of global influencer marketing dollars.

Let’s see how it breaks down.

The Backstory

Influencer marketing is enormous business (estimated between USD $3.2 and 6.3 billion in 2018). Clearly, this entire economy would be impossible without Facebook and Google’s platforms. Instagram and YouTube are the lead platforms for influencer marketing worldwide. But until recently, Facebook wasn’t seeing a sliver of this value enter their bank accounts.

Meanwhile, Google is keeping 45% of every dollar spent on ads bought against YouTubers’ content. Through this ad-based video monetization and other means on YouTube, Google has been making money off influencers for years.

It’s only natural that Facebook would want its slice of that pie as well. We’ve already seen previous moves in this direction.

The Branded Partner Feature

First, in June of 2017 Facebook launched the Branded Partner Feature. This allows influencers to tag a post as sponsored by a particular brand. If an influencer does so, the brand can boost that post with paid support. They can do so by boosting the influencer’s post, without sharing the post on their own page first. (I wrote more about this last August).

Tagging of this type is necessary for regulatory compliance, as FTC and ASA rules require influencer content to be clearly identified as sponsored.

But this also means Facebook knows which posts by an influencer are organic and which are sponsored by a brand. So Facebook could – in theory – strangle organic reach to zero for every sponsored post and demand payment from the sponsoring brand.

After this, Facebook updated their algorithm to prioritize posts by people over brands, raising the relative importance of influencer content to marketers.

Note that Facebook hasn’t yet strangled organic reach on sponsored posts, on FB itself or Instagram. But their track record suggests they are likely to do so – and it makes great business sense.

Enter the Creator App

Now that brands have an incentive to work with creators on the Facebook platform, it’s important for Facebook to ensure that these advertisers have lots of great inventory to work with.

So what’s important to brands?

A vast variety of fresh, high-quality content to sponsor.

That’s why the Facebook Creator App exists. It’s designed to make it easier for creators to create content, manage their audience interactions and content performance.

There’s no explicit creator monetization features for now – but that might be coming. It’s also less important in light of what Facebook announced last week.

Then, Branded Content Matching

Last week Facebook announced a third development to its platform, which increases Facebook’s offer to brands who want to use the platform for influencer marketing.

The Branded Content Matching feature allows brands to locate influencers on the platform based on audience demographics and biographical features. You can use the BCM feature to find influencers with audiences of mostly married people who own houses, for example, or people in a particular city with kids.

Needless to say this is very powerful audience-based segmentation. Exactly what you’d want if you were a brand trying to choose a micro-influencer to work with, based on ROI rather than lifetime follower counts.

Is this the end for the micro-influencer platform agencies?

Because of this change, a whole crop of businesses is potentially going to die.

Micro-influencer platforms have been mushrooming lately as canny entrepreneurs rushed to fill the influencer marketing niche. Platforms like Indahash, TRIBE and others live to match brands with micro-influencers who can feature their products for a fee. The way these platforms work is that influencers have to choose to sign up, and then brands can give them briefs through the platform. Some of these businesses are so specialized that they exist specifically for Instagram influencers.

But as I’ve said for a while, when it comes to micro-influencer platforms there can be only one. Ultimately the market economics of this work according to network effects – any platform with more influencers signed up will tend to attract more influencers, and then more brands, in a self-reinforcing cycle. The money comes from brands, and they are always going to plump for the platform with the biggest inventory. And Facebook has the biggest inventory of all.

So with this move, Facebook could potentially crush all the smaller players in the field.

So where does this leave us?

As marketers, this is good for us. In a way.

Influencer marketing has been a sort of Wild West free-for-all for years. From a marketer’s point of view there has been a confusing profusion of intermediaries, agencies and startups. Influencers can’t even agree among themselves how much they should be paid for sponsored content.

With this change, Facebook is consolidating and simplifying the marketplace. They’ve assembled in one place all the tools needed really effectively find, engage and manage influencers on Facebook and Instagram – two of the main platforms we tend to work on.

On the other hand, this is all going to come at a cost. If I were in charge at FB, I’d start gradually but inexorably lowering organic reach on sponsored content to zero, and making big brands fork over big money to reach those specialized, high-ROI audiences.

We need to get ready for influencer marketing to become more formalized, and more expensive.