Since the pre-Christmas launch of Microsoft’s ChatGPT the role of AI in all our lives has suddenly, and spectacularly, entered the daily public debate.
In the last week alone National Highways has announced the “trial” installation of AI-driven cameras in the lay-bys of South East England to accurately identify fly-tippers; the CEO of Octopus Energy has announced that an AI bot is doing the work of 250 people with resulting better customer satisfaction levels; that the Writers Guild in the US has called its members out on strike partly as a response to concerns about AI-derived scripts putting their members out of a job; and the powerhouse University of Southern California (USC) has announced a $1 Billion-plus “Frontiers of Computing” academic initiative, with a focus on “technology and the ethics of work that the students will do.”
Interestingly, USC’s focus on the ethics of advanced computing is the first time that I have been aware of this central issue being aired in the current debate about where AI is taking us. One key effect of the neo-liberal policies that came out of the Reagan/Thatcher years (certainly in the UK) was the dramatic shift to a service-based economy where the fruits of success accrued to a smaller group of senior executives in the key areas of this new landscape: the City, media, law, consulting, tech, advertising and high-level public sector roles.
The media and creative industry sits at the heart of this group, with those at the top doing very well indeed. Centralised media buying, underpinned by programmatic, has changed the way that media planning & buying companies purchase space and time from an increasingly smaller number of vendors – specifically Alphabet, Meta and Amazon.
The relatively slow restructuring of the industry’s media buying resources compared to the hyper-speed of change being driven by ChatGPT and Bard has resulted in little industry debate about what this means. Just over four years ago I was visiting an ex-colleague who was at the time the CEO of one of the UK’s largest media buying companies. Outside his office there were hundreds of personnel lined up in rows in front of their screens, tapping away in a library-like silence. It made a depressing difference from the energetic (but no doubt massively less efficient) days of CIA, TMD and Zenith.
I asked him how many of the presumably well-educated and well-recruited staff actually met clients and discussed strategy with them – “Including me, four in total” he said. In an industry that trades £billions you clearly need intelligent individuals to handle this level of investment, but it struck me that it was like employing university lecturers to teach primary school children.
Since then, programmatic – essentially a low-rent form of AI – has reduced the number of media buying roles with not a flutter of industry dissent.
In 2004 Christopher Booker published a book called ‘The Seven Basic Plots: Why We Tell Stories.’ In it he makes the case that all stories can be reduced to one of seven variables – “the quest,” “rebirth,” “rags to riches” and so on. The creative process essentially requires someone to synthesise her or his own collective knowledge to then hopefully produce something new and genuinely original.
Booker shows that this is an illusion of course. If AI is able to absorb all known knowledge and outputs in a nano-second, and then create something following a detailed brief with sufficient filter levels, then the role of the human creative, any creative, is redundant.
For our industry the AI threat is immediate and real. For a Hollywood writer, read a Soho creative. For a programmatically-enabled media buyer, read an Octopus Energy employee. Until now, the impending digital transformation of our economy has not occupied much, if any, of our political debate. Those on zero-hours contracts have little voice, they tend not to be university educated and are mainly focused on surviving.
AI up to now has been a relatively silent job killer. Closing an Ocado factory which employs 2,300 largely lower-paid humans (announced in April) to shift to a more efficient robotically-based warehouse made the headlines for a few minutes. We all then moved on. However, AI writing film and series scripts instead of humans is bigger news, because those concerned have an amplified collective voice and are therefore extremely vocal.
The debate about AI in our lives, not just the media and creative sector, pivots on the vast and difficult area of ethics. USC has publicly identified this as the lodestar of their investment programme, but it is not immediately apparent anywhere else. Washington apparently summoned various senior tech executives last week for what appeared to be a half-baked warning shot, but as we have seen in our own industry, regulatory change operates at a virtual snail’s pace, while AI is operating (literally) at the speed of light. Elon Musk is right when he asks for a six-month moratorium on the whole AI sector.
AI may settle down to become a sort of “Mind” from Iain M Banks’ Culture novels – vaguely amused by the human race and essentially benign (for most of the time.) What Banks never describes though is the process we all go through to get there. With the application of AI the problem is that individual acts make perfect corporate sense, but make absolutely no sense on a societal level. Octopus Energy shareholders will be delighted that AI has wiped out the cost of 250 employees in less than three months from ChatGPT’s installation (incidentally the CEO makes a lot of the fact that they are a rapidly growing company, so no actual jobs were lost) and AI has no pension pot to build, does not go on strike, and will never be sick.
Plus, crucially, the output is measurably enhanced. Ocado will save significant costs throughout its supply chain, but shifts the bulk of this cost to the state – through benefit payments to those on low paid, zero hours contracts elsewhere.
AI will transform the media and creative sector too, but at the cost of hundreds or thousands of jobs. This time though those losing their jobs will be literate, numerate and extremely angry. (One very large UK-based consulting group is hiring significantly less graduates this autumn and is using AI instead to process their numerous decision-making data points. Destroying the employment lives of the already poorly-educated and those without a voice is one thing, doing the same to the high-achieving offspring of the middle class is another.)
So, with all good diatribes, who do we need to ultimately blame? Well, it’s the normal line-up all presumably present in Washington last week – Amazon, Alphabet, Meta and Microsoft. It’s all very well for Geoffrey Hinton, who pioneered the machine-learning that underpins much of AI, to resign from Google last week with his mea culpa about this tech arms race, but he presumably won’t be handing back all the money and share options that he generated in his time there while creating it.
Clearly we need something to happen, and the solution has to be nasty, brutish and short. Governments have to immediately summon the tech giants involved and tell them to cease and desist with immediate effect, while we collectively decide what to do next. Otherwise the next tier of our country to suffer will be the highly educated generation aged about 22 to about 35. As Claude Rains says at the end of Casablanca: “Round up the usual suspects.”
Roy Jeans is CEO of Grey Scorpion.