At Fifty Five and Five, we’ve witnessed first hand how technology reshapes our industry. And as we all know AI is next level stuff. I remember when we first demoed our in house AI tool Compass to a client —it felt like unlocking hidden potential. Or when we updated our writing process to use AI, allowing anyone in the company (not just our writers) to create professional grade copy. We have whole new ways of working that simply were not possible even just months ago. AI can even draw Pelicans!
However, our excitement at AI comes with pragmatic caution. And now this. Meta (Facebook’s parent) is betting big on AI, claiming it can plan, create, and optimise ads better than any human agency. Mark Zuckerberg’s bold vision is sparking a mixture of hope and scepticism about the future of advertising. Is this the dawn of a revolution where businesses simply plug in their budget and let Meta’s AI do the rest? Did anyone ask for that? Should advertisers and agencies in the UK and beyond be wary of handing over the keys?
This blog explores Meta’s AI advertising vision, how it works today, and my take on whether we should embrace it or hit the brakes (Brakes! It is always the brakes!)
Zuckerberg’s vision: Meta’s AI as your ad agency
Meta’s CEO Mark Zuckerberg envisions an AI-driven ad platform that replaces many functions of an ad agency. In an interview this week, he described a future where a business owner provides only a product, an objective, and a bank account – and Meta’s AI handles everything else.
Creative content (images, videos, copywriting) would be generated by AI, targeting and placement optimised by AI, and performance measured and iterated by AI, all within Meta’s ecosystem. Zuckerberg suggested that eventually “you don’t need any creative, you don’t need any targeting... just connect to your bank account” and read the results the AI.
“Just connect your bank account”! Was there ever a more blatant statement of naked corporate greed? Sorry, let's get back to being objective.
This move essentially positions Meta’s AI as a one-stop advertising agency, handling campaign strategy through to sales. It’s a sweeping vision to “redefine the category of advertising” – one that could eliminate traditional ad creation and media buying roles. Zuckerberg claims such AI automation will make advertising vastly more efficient and even expand the overall advertising pie (he predicted AI will make advertising a “meaningfully larger share of global GDP” over time).
The core promise is sort of seductive: infinite personalised ads at scale, minimal effort from businesses, and potentially better results. Look we know from our experience at Fifty Five and Five that AI massively helps in this world. For instance, our AI marketing platform Compass Agents helped one client increase their ad engagement rates by 40% last quarter by automating dynamic content adjustments.
But is Meta’s AI anywhere near capable of doing what they claim? And way way more importantly do we even want it to be?
Meta’s AI ad platform: How it works today (Features & limits)
In practice, Meta’s AI advertising tools are already rolling out – but they’re not magic. Currently, Meta offers generative AI features in its Ads Manager that assist with creative production and optimisation. For example, tools like Background Generation, Image Expansion, and Text Variations allow advertisers to automatically create new ad visuals and copy based on existing assets. An advertiser can supply a product image, and the AI will generate alternate backgrounds or even entirely new image variations of the scene. It can also suggest multiple versions of ad text (headlines or captions) to test different messaging. These features help produce the “infinite creative” Zuckerberg touted, by churning out many ad versions to see what works.
Importantly, these AI tools are integrated into Meta’s standard ad products (like Advantage+ campaigns). For instance, Meta’s Advantage+ shopping campaigns have an option to auto-generate creative – a case study saw a 13% boost in return on ad spend when a retailer used the GenAI background tool. Meta has also introduced AI-driven targeting and budgeting optimisations under its Advantage suite, where advertisers set a goal and the algorithm finds the best audience and spend allocation over Facebook, Instagram, etc. In short, Meta’s platform is moving towards the “give us your goal, we’ll do the rest” model.
However, limitations still apply. Advertisers currently must provide some base inputs (imagery or text) for the AI to work with – it’s not yet creating entire campaigns from nothing. The AI can remix and augment content, but truly autonomous ad creation (“just give objective and budget”) is still in pilot phases. Meta is cautious with sensitive sectors too: it has barred political, healthcare, finance, and other regulated advertisers from using these generative tools for now, to avoid misuse (e.g. fake or misleading imagery in election ads). The AI also isn’t infallible – it can generate low-quality or off-brand outputs if not guided, so human oversight remains important. Guardrails are in place (Meta says it filters out inappropriate or unrealistic AI content), but brands must still vet the AI’s creations for accuracy (e.g. ensuring an AI-generated image doesn’t show a product in a colour or form that doesn’t exist). In summary, Meta’s AI ad platform today is a powerful assistant for producing and optimising ads, but it’s not yet a fully hands-free ad agency. Businesses can experiment with AI-generated creatives and auto-targeting, while still steering brand messaging and ensuring nothing goes off-brand or illegal.
Advertiser reception: Enthusiasm vs. scepticism
The advertising community’s response to Meta’s AI push has been mixed. On one hand, adoption of Meta’s AI tools has been massive among advertisers, especially small and mid-sized businesses hungry for easy creative solutions. By late 2024, over 4 million advertisers – up from 1 million just six months prior – had used at least one generative AI tool in Meta’s ad platform. Meta boasts that in a single month, more than 15 million ads were created with AI. These users are lured by promising stats: campaigns using Meta’s generative AI creatives saw an average 11% higher click-through rate and 7.6% higher conversions in tests about. For many resource-strapped advertisers, especially startups and small businesses, the idea of pumping out lots of tailored ads without hiring a big agency is highly appealing. Success stories are emerging – e.g. Casetify, a phone case maker, reported a 13% uptick in ROAS (return on ad spend) after trying Meta’s AI Background Generation in its campaigns. This has created a sense of optimism that AI can deliver better performance with less effort.
On the other hand, seasoned marketers and large brand advertisers are more cautious. I am way more cautious! Fifty Five and Five is, and we are all in on AI generally!
Sure we can acknowledge the potential efficiency gains but we recommend client don’t entrust their brand entirely to Meta’s black-box AI. We worry an algorithm might produce content or placements that don’t align with brand values or run in inappropriate contexts – problems agencies like us normally guard against. For large clients used to rigorous campaign planning, A/B testing, and independent measurement, the idea of just “reading the results [Meta] spits out” feels like checking Meta’s homework with no transparency. Trust is a hurdle: we remember past measurement scandals and ad fraud issues in digital ads all too well, so we are reluctant to blindly accept performance reports from the very platform spending their own clients money.
Risks and criticisms: Brand safety, creative control & Meta’s dominance
Handing the reins to Meta’s AI raises several red flags for me. Big red wavy flags. Brand safety is one major concern. If Meta’s AI is auto-generating countless ads, can my clients be sure none of those ads inadvertently harm the brand’s image or appear in the wrong context? Meta’s track record on content moderation isn’t spotless, and I fear scenarios where AI-created content might be off-tone or even offensive. Me and my team using AI to augment content and copy, with guardrails and approvals is one thing. Meta doing it as fast and cheap as it can to maximise returns is a very different one.
Plus risk of homogenisation is real too – if every advertiser uses the same AI, creative outputs could become formulaic. Brands risk losing their distinct voice in a sea of AI-generated, cookie-cutter ads. We already see this with AI content slop. I am a massive believer that AI can be creative in its own right, but maybe not just yet.
Control is another sticking point as far as I see. Companies invest heavily in brand identity and messaging; surrendering creative decisions to Meta’s algorithms can feel like ceding the very storytelling that defines a brand. An AI might prioritise what gets clicks in the short term, but does it understand a brand’s nuanced values and long-term equity? We do.
I fear, as do our clients, “frankenstein” content that, while optimised for performance, doesn’t feel authentic. Even Meta acknowledges some pitfalls – for instance, its image AI could hypothetically show products in forms that don’t exist, which could mislead customers or dilute trust. Such scenarios underscore why myself and many marketers insist on human review and veto power over AI suggestions. In regulated industries, these concerns double – hence Meta’s preemptive ban on using generative AI for credit, housing, political and pharma ads. The last thing a brand needs is an AI-generated message that runs afoul of laws or public sensitivities.
Then there’s Meta’s dominant role. If Meta’s vision comes to pass, the company wouldn’t just be the media platform, it would also become the creative agency, the media buyer, and the analytics provider all in one. This end-to-end dominance raises conflict of interest and transparency issues.
Meta essentially would be “marking its own homework” – providing the performance metrics on campaigns that its own AI created and ran.
Obviously companies like Fifty Five and Five want this work. But also we worry about a lack of independent oversight: Years of audits and third-party verification in advertising have taught us to verify platform-reported data. If Meta’s AI agency model bypasses independent measurement, clients could be flying blind, uncertain if the results are as good as claimed or if spend could be optimised better elsewhere.
Moreover, Meta’s profit motives might not always align perfectly with our clients interests – an AI optimising for conversions might also subtly encourage higher spending on Meta’s network. Without external agencies or auditors in the mix, who ensures Meta’s AI isn’t, say, favouring Meta’s revenues over the advertiser’s ROI? These dominance concerns aren’t just paranoia; they echo real incidents (like past misreporting of video metrics by social platforms).
Meta’s motivations: Profits, power, and “Infinite” ad Spend
Why is Meta so keen on having its AI take over your advertising? In a word: money. Advertising is Meta’s cash cow, and anything that makes advertising on Meta easier and more effective stands to hugely benefit Meta’s bottom line. Zuckerberg himself has argued that AI will let many more businesses advertise, increasing the total advertising spend globally (hence his comment that it could boost ads’ share of GDP).
By lowering the skill barrier – “all you need is a product and a budget” – Meta could unlock a flood of small and medium businesses who previously couldn’t afford professional ad services like ours.
If millions of businesses hand Meta their bank account to automate ads, Meta captures not only the advertising spend but also potentially budgets that might have gone to creative agencies like Fifty Five and Five. This disintermediation means a larger slice of the marketing pie for Meta itself. It’s telling that Meta’s AI tools are free or low-cost to use – the company knows it will earn back its investment through increased ad volume and spend.
There’s also a lock-in effect. If Meta’s AI handles everything, advertisers might rely exclusively on Meta’s platforms (Facebook, Instagram, etc.) for their marketing, at the expense of other channels. This concentrates spend within Meta’s walled garden. Over time, Meta could gain unprecedented influence over how marketing budgets are allocated. Consider that if an AI is choosing where to show your ads, it’s naturally going to stick to Meta-owned properties. That could subtly divert money from competitors (like Google, TikTok, or traditional media) into Meta’s pocket. Meta essentially becomes the gatekeeper of your customer acquisition. The more it can claim its AI yields the best results, the more marketers will double down on Meta – a virtuous cycle for Meta’s finances (but potentially a vicious cycle for market competition).
Meta also has a data advantage: by running everything through its AI, it collects even more granular data on what creative and targeting works. This can further improve its algorithms, widening the gap between Meta and any outside ad tools. Smaller agencies like us simply can’t match the scale of data Meta’s AI is trained on. Over time, this could make Meta’s optimisation so “smart” that advertisers feel they have no choice but to use it – again boosting Meta’s position. Financially, Meta saves on having to court big agencies or third-party ad tech integrations; instead, budgets flow directly. And if results are good, advertisers might even increase spend (CFO Susan Li noted Meta’s cost-per-ad is rising as AI drives more, indicating advertisers value the outcomes enough to pay more).
However, it’s worth noting Meta’s AI push is also defensive, they are not alone. Competitors like Google are developing similar AI ad tools, and the generative AI wave threatens to upend how marketing is done across the board. Meta doesn’t want to be disinter-mediated by outside AI solutions, so it’s building its own to keep advertisers within its ecosystem.
By being the first to offer a full-suite AI ad service, Meta aims to capture the narrative (and cash) early. The company’s near-unilateral control by Zuckerberg (who holds ~54% voting power despite ~13% share ownership) means it can pursue this aggressive strategy without shareholder interference. As a result, Meta stands to benefit disproportionately from an AI-centric ad future – reaping more spend and consolidating power – which is exactly why advertisers should think carefully about how much control they concede. The financial motive isn’t evil per se (every company wants growth, we do!), but here it aligns with reducing checks and balances in the ad system. That imbalance of power and profit is at the heart of why this “revolution” deserves a sober examination.
The Zuckerberg factor: Trust and control in one man’s hands
Another aspect I think people should look at is the unique control Mark Zuckerberg wields over Meta. Unlike many public companies, Meta is essentially Zuckerberg’s kingdom – through a special stock structure often likened to a “golden share”, he can’t be easily overruled by other shareholders. In practical terms, this means if Zuckerberg believes Meta’s future is an AI-run ad platform, he can charge ahead, even if some partners or investors have misgivings.
You will be well aware that Meta’s leadership has a history of unilateral big bets (e.g. the pivot to the metaverse), and that external voices have limited ability to change Meta’s. This cuts both ways: on one hand, Meta can innovate quickly and ambitiously (as seen with its heavy investment in AI), but on the other, it can make sweeping changes without the consensus-building that might happen in a more accountable corporate governance setup.
Do we really trust in Meta’s stewardship? Zuckerberg’s power concentration raises questions: Will Meta’s AI always act in clients’ best interests, or will it serve Meta’s strategic goals first? If problems arise – say widespread AI-generated ad mistakes or controversies – do advertisers have any recourse or influence to demand fixes?
Feels fishy to me.
My concern is that with Meta marking its own homework and the CEO holding all the cards, agencies like us might be left in the dark or forced to accept Meta’s assurances. For instance, if Meta’s AI reports great results, but an advertiser wants an audit or an explanation, Meta doesn’t have to oblige in a world where they run the show end-to-end. This dynamic makes some advertisers uneasy; it requires a lot of trust to give one company so much control over your marketing, especially when that company is effectively controlled by one individual’s vision.
Agency response: Adaptation, resistance, or partnership?
Advertising agencies like Fifty Five and Five are directly in the crosshairs of Meta’s AI ambitions. If Meta’s AI can truly do it all – strategy, creative, media buying – what’s left for agencies? I see a spectrum of responses from resistance to adaptation to partnership as agencies navigate this disruption.
Resistance: Some agencies are pushing back subtly by highlighting the continued importance of human creativity and independent strategy. They argue that while AI can generate content, it lacks the human insight into culture, emotion, and storytelling that builds iconic brands. Agencies are reminding clients that an algorithm might optimise for clicks, but a big creative idea can ignite a brand love that metrics can’t immediately capture. In pitches and industry forums, you’ll hear the refrain that “AI can’t replace the creative magic” – an attempt to defend the agency’s role. Additionally, agencies stress their ability to work across platforms. A Meta-run campaign will inevitably favour Meta’s channels, but an independent media agency can balance spend between Meta, Google, TikTok, TV, etc., to suit the client’s best interests. This multi-channel stewardship is a key value prop agencies highlight to dissuade clients from going all-in with Meta. Some are even cautioning that giving one platform too much control could be risky – essentially telling advertisers, “don’t put all your eggs in one basket, especially one you don’t fully control.”
Adaptation: Many agencies are also pragmatically adapting by embedding AI into their own workflows. Rather than see AI as the enemy, they are using it as a tool to enhance productivity. Creative agencies, for example, use generative AI (from various providers) to produce first drafts or multiple variations of campaign ideas, which human creatives then refine. Media agencies are deploying AI for better data analysis, budget pacing, and even to auto-generate reports. The idea is to augment human teams with AI, so they can deliver results as efficiently as any in-house Meta system, but with an added layer of human judgment and cross-platform perspective. Agencies are also investing in proprietary AI solutions. Notably, WPP (a major UK-based agency holding group) partnered with NVIDIA to build a generative AI “content engine”** for ad production. This tool aims to let WPP quickly create high-quality ad content at scale using AI, keeping the capability in-house rather than ceding it to Meta. Such moves indicate that agencies plan to compete on AI – they want to tell clients, “We have AI superpowers too, and ours work on any platform, not just Meta.”
Partnership: Interestingly, some agencies are choosing to work with Meta on this journey. Meta has been actively engaging agencies through advisory councils and pilot programmes (even sending engineers to agencies to collaborate on AI ad tool development). This partnership approach suggests agencies can have a role in shaping Meta’s tools and perhaps offering value-added services on top of them. For instance, an agency might specialise in crafting the best prompts or feeding the right brand inputs to Meta’s AI, effectively acting as a “meta-whisperer” on behalf of clients. In the UK, big agencies and Meta have historically had a symbiotic relationship – agencies bring the big budgets, Meta provides the reach. That dynamic is evolving, but not necessarily collapsing. Smart agencies are looking at consultative roles, helping clients navigate when to use AI, how to maintain brand consistency, and how to interpret the results Meta’s systems provide. They may also double down on areas AI currently can’t do well: high-level brand strategy, experiential marketing, or big creative stunts that cut through the AI noise.
Fifty Five and Five is doing a combination of all three. We are pushing back on needing humans and independent strategy. 100%. But we are also all in on the right AI tools and techniques. So it is a blend. We also believe that AI will be doing more and more as it gets better and better. And yes of course we partner with Meta, we use their tools and platforms. We can’t not.
UK perspective: Cautious optimism and scrutiny
The issues above carry particular weight in the UK’s advertising industry, which is one of the most advanced and brand-sensitive markets in the world. British advertisers and regulators have been notably vocal about transparency, trust, and responsibility in digital advertising – values that come sharply into play with AI-generated ads. UK brands remember well the “brand safety” crises of recent years (like ads appearing next to extremist content), and they have led global pushes for stricter standards. It’s no surprise that in the UK, industry bodies like ISBA (representing advertisers) are actively examining generative AI. In fact, ISBA has emphasised the need for clear transparency when AI is used in ad production, and Meta has responded by agreeing to label AI-crafted ads in its platforms. This shows the UK’s influence in demanding safeguards – any AI revolution in advertising here will be expected to meet high standards of disclosure and brand protection.
Moreover, the UK’s regulatory environment is gearing up for Big Tech oversight. The Competition and Markets Authority (CMA) has investigated the digital advertising duopoly of Meta and Google, and a new Digital Markets Unit is in the works to ensure fair play. If Meta’s AI agency model were to significantly increase its dominance, you can bet UK regulators would scrutinise it for anti-competitive implications. The UK ad market values competition and plurality – agencies, media owners, and advertisers all coexist. A scenario where Meta gobbles up more of the value chain could trigger policy responses (for example, mandates on data transparency or interoperability of ad tools). Even on data privacy, the UK (and EU) have stricter laws than the US, meaning Meta’s AI will be under pressure to handle data and personalisation in compliant ways. Advertisers here will ask: how is the AI targeting users? Is it respecting consent and privacy rules? These questions are front-of-mind in Europe, and any misstep could erode trust or invite regulatory action.
That said, the UK advertising community also sees opportunity. London is a global hub for ad tech innovation and creative excellence. Many UK-based marketers are experimenting with AI – but often with a healthy dose of scepticism. The attitude is often “trust, but verify.” UK brands might try Meta’s AI tools on a small scale to gauge results, but they’re less likely to hand over full control without rigorous testing. There’s also a push for up skilling – the IPA (Institute of Practitioners in Advertising) and other groups are training agency staff on AI, ensuring the human talent remains knowledgeable and can question or complement the AI. And of course, British agencies like WPP are at the forefront of developing independent AI solutions, proving that the innovation race is not solely in Silicon Valley’s hands.
In summary, the UK relevance here is in the balanced approach, like that of Fifty Five and Five. Embracing the efficiency and creative boost AI can provide, but with strong demands for trust, accountability, and competition. The question “Should we hand over control to Meta?” might get a more cautious answer in the UK than elsewhere. British advertisers are likely to insist on a partnership approach with Meta’s AI – leveraging its strengths but also keeping their own agency partners in the loop, insisting on transparency, and holding Meta to high standards. The UK’s legacy of advertising effectiveness and self-regulation means the hype will be tempered with pragmatism. It’s an important bellwether: if Meta’s AI can win over the cautious, scrutinising UK market, it bodes well – but if not, it’s a sign that more work is needed to build trust.
Conclusion: Hand over control – or proceed with caution?
Meta’s AI advertising revolution offers a tantalising promise FOR SOME: faster, cheaper, and smarter ads generated at the click of a button, potentially levelling the playing field for businesses of all sizes. Mark Zuckerberg’s vision of an AI “ad agency” in a box could indeed transform how marketing is done, and early results show real upside in performance. However, as we’ve explored, handing over full control to Meta is not a decision to be taken lightly. The efficiencies come bundled with significant concerns about oversight, transparency, and dependency on one dominant platform. Advertisers must weigh the allure of automation against the loss of independent creative control and the risks of trusting Meta to grade its own homework.
For now, a hybrid approach seems wise. We will embrace Meta’s AI tools to augment their work – use them to generate ideas, streamline production, and improve targeting – but keep a human in the loop. Treat Meta’s AI as a powerful assistant, not a total replacement. Insist on seeing the data and understanding the logic behind the AI’s decisions (as much as possible). Continue using third-party measurement or your own analytics to verify results, so you’re not solely reliant on Meta’s reporting.
In other words, leverage the “revolution” without becoming wholly captive to it. Agencies should continue to play the roles of strategists, auditors, and brand stewards in tandem with AI. And Meta, for its part, will need to continue earning trust – through transparency measures (like labelling AI content, opening up about how its models work, allowing audits) and respecting that advertisers deserve agency (both the noun and the verb).
The question “Should we hand over control to Meta?” is easy. It’s a ‘No’. There’s little doubt AI will play a big role in the future of advertising – the revolution is already underway. That’s fine. But perhaps the healthiest outcome is one where Meta’s AI is a tool in the hands of marketers and agencies, rather than a black box we blindly hand the keys to. In the UK and globally, those who strike the right balance – embracing innovation while demanding accountability – will likely reap the rewards of AI without losing sight of the creative and strategic human touch that truly builds brands. The revolution in advertising is here, but it’s up to us to shape it, not just ride it.
