Published in May-Jun 2023
“History never repeats itself, but it does often rhyme,” said Mark Twain. AI has triggered a predictable flurry of techno optimism and pessimism. But what might it mean for communications? This little bit of futurology is the second round for me. Twenty years ago, I was researching a book predicting the transformative impact of digital technology. I attended many conferences and had fascinating meetings with techno visionaries. I got some things right and some wrong. What did I learn from that experience? Here are my predictions:
1. Predictions about what might happen in more than five years are guesswork: In 2004 it was obvious that media supply would explode with almost limitless numbers of websites as well as many more TV channels. The age of total media was arriving fast. Almost anything could be an advertising medium, and many media would be interactive allowing people to talk back. So far, so bland. What happened next was this. A combination of ‘enabling technologies’ provided the platform for something no one predicted. Broadband, 4G, ever faster processing speeds, cloud computing. Then came the innovators – Google and Facebook – which consumed the business models of the legacy media, especially the newspapers, which lost their classified listings revenue in short order. Then came the smartphone, supported by GPS, and the App economy which triggered a boom in service innovation. Walk down any street now and you will see people prodding and gazing at a small screen and occasionally talking out loud, seemingly to no one in particular. They are carrying the tiny super computers we now call smartphones. If I had written those last two sentences 20 years ago, it would have been a work of science fiction. Science fiction may be the richest way to conjure up a picture of the distant future. The creators of Star Trek got closer to imagining the smartphone than any number of well-informed futurists.
2. Beware of Endism: The arrival of a new technology provokes futurists into predicting the death of the old. Thus, the internet was predicted to kill TV advertising. TV ads did not die because a well-made short film is a pleasanter experience than being interrupted with a message, however well targeted. Today AI is predicted to kill search – Googlers are fretting that they have “no moat”, no protection against competition.
3. Search will not end. It will get better: Those Googlers are worried because they don’t yet know how to make money out of AI and fear it will kill the cash machine of paid search. Unlikely. Firstly, for human reasons. Googling is an ingrained habit which will not change quickly. Secondly, Google’s new AI-driven service called Bard is complementary to search. For example, I have just searched “best restaurants in Soho” using both Google and Bard (this being one of the few topics I have expertise in). Google delivers pages of links to reviews, ads and restaurant websites – useful to find trusted recommenders and to make a booking. Bard recommends the five best. It is a credible list; useful if I am short of time and don’t want to trawl through websites. I expect Alphabet to integrate the two services. They may be slow about it – dominant brands are often conservative – as they don’t want to eat their own lunch. (Declaration of interest. I am an ex-Googler: I left five years ago. I also predict that Google will lay off more staff, stop wasting money on “moon-shot projects” and focus investment on the existential threat of AI.)
4. Regulation Will Be Coming: Bad people are very innovative and will use AI for nefarious purposes. In retrospect, we were much too starry-eyed about the digital revolution 20 years ago. Something as huge as the internet was always going to reflect all of human nature, including the mendacious, evil and greedy. It took governments at least a decade to develop regulations. This time, governments will be quicker off the mark. China and Europe will lead in their own ways. Bill Clinton predicted that the Chinese would find it impossible to control the internet (“Like nailing Jello to a wall,” he said). He was wrong. The Chinese Communist Party always invests big in any neutralising threats to its authority, even to the extent of undermining its own tech companies. The EU worries that it will be overwhelmed by US tech giants and will regulate first. The US will have the lightest touch regulation to ensure that their big tech continues to dominate. The UK will follow the US to try to develop a strong position in an emerging technology.
5. AI will get cheaper quickly and many will benefit: A brief explanation for non-techies. By AI we normally mean “Large Language Models (LLMs) which require huge amount of data, big computing power and time. Expensive. But, as The Economist explains, ‘LLMs can be fine-tuned using a technique called low-rank adoption… an existing LLM can be optimised for a particular task quickly and cheaply.’” Something similar happened with websites. Very expensive 20 years ago, as they were designed and coded from scratch, then free blogging took off, which at first looked amateur and crude. Now you can design and maintain a professional-looking e-commerce site using preformatted elements and plug-ins for just a few hundred dollars. Businesses large and small will enjoy both service improvements and efficiency gains using optimised LLMs dedicated to a range of applications, delivering customer services via chatbots (this is already happening), summarising the issues, writing a draft plan, generating and optimising your response marketing campaign, writing code. Much more besides. You can already train your own AI companion/partner using an app called Replika. Customers report that they are often more empathetic than the flesh and blood equivalent.
6. There will be a backlash that brands will need to understand: Rapid change always creates a feeling of being unmoored. It is a story as old as the Industrial Revolution. Anxiety about AI in marketing will be high because it challenges what we understand by a brand. A brand is created and managed by people. Brands have ‘values’ and ‘purpose,’ which are expressions of human morality; of knowing the difference between right and wrong. AI has no morality. It will give you plausible answers and (increasingly) expert answers. But it can’t say: I have looked at all the evidence and in spite of this, I think we should do something different because it is the right thing to do. Companies will use AI to make efficiency gains and cut costs, but they will need to re-invest in ‘human capital.’ Call centres may be replaced by chatbots but better-trained people will always need to be on hand. Brands will use AI to automate mundane and repetitive tasks but the winning brands will be those that give you the option to speak to experts.
A few words of advice. It is difficult to predict technological change, but if you keep human nature to the fore, as I have tried to do in this article, you should be able to navigate the next few years. It is also a good idea to put some budget aside to experiment and learn. The good news is – it will be affordable quite soon.
Julian Saunders is a former ad agency CEO and Googler who has worked for the UK government. firstname.lastname@example.org