ai

Empire of AI by Karen Hao

Peter Cripps review , , , , , ,

Karen Hao’s Empire of AI is without doubt a significant contribution to the ever growing collection of books digging into the workings of the AI industry. It is not just an insiders exposé of the shenanigans that went on leading up to the firing (and rapid rehiring) of Sam Altman, OpenAI’s founder and CEO, but also provides a deeply researched and incredibly well written insight into the current state of play of the AI industry as a whole, and, it’s not pretty.

The book opens with the events that took place over the long weekend of Friday 17th November to Tuesday 21st November 2023. On Friday 17th, Altman was invited to a video call set up by the OpenAI board members where he was unceremoniously told he was being fired. However by the following Tuesday, due to overwhelming pressure from OpenAI employees and, more importantly its investors, especially Microsoft, Altman was back at the helm with a new board of directors. Empire of AI examines how Altman and his fellow conspirators came to create and dominate the techno-industrial complex that is referred to generically as ‘AI’ and how, if things carry on as they currently are, we risk destroying jobs, the environment and most, if not all, forms of human endeavour and behaviour.

Empire of AI is divided into four parts. Part I covers how Sam (Altman) met Elon (Musk), the latter being a “childhood hero” of the former, and decided to build an AI company that would compete with Google head-on to beat them in the race to building AGI (artificial general intelligence). Musk was fearful that Google’s takeover of the British AI research company, DeepMind would lead them to develop AGI first and thereafter “murder all competing AI researchers“. Musk was adamant that “the future of AI should not be controlled by Larry (Page)“. To Musk, Altman was a like minded entrepreneur who wanted AGI to be for the good of all humanity, not something that would allow Google to become even richer by destroying all competitors in its wake. OpenAI was formed with good intentions therefore, to create the first AGI that could be “used for individual empowerment” and which would have safety as “a first-class-requirement“.

OpenAI was launched as a nonprofit company in December 2015. Altman’s co-founders were Greg Brockman (an engineer and fellow entrepreneur) and Ilya Sutskever (an AI researcher poached from Google). The company had a $1B commitment from, amongst others Elon Musk, Peter Thiel (Musk’s fellow cofounder of PayPal) and Reid Hoffman (another of the so called “PayPal mafia” and cofounder of LinkedIn). Having started a company with the ultimate goal of developing AGI, OpenAI needed to do three things quickly – figure out exactly what it was they were going to build to achieve AGI, hire the right talent to do this whilst at the same time securing enough funding to make these two things possible. By 2019 these problems seemed to have been solved. The company would focus on building its AI technology using a large language model (LLM) it called GPT-2. In order to secure the necessary funding to be able to pay for the enormous amounts of compute LLMs needed they switched from being a nonprofit to a for-profit company opening up the floodgates to companies like Microsoft investing in them in the hope of making huge profits if OpenAI were the first to achieve AGI. On this basis Microsoft announced in July 2019 it was to invest $1B in the company.

In Part II the book looks at some of the looming problems OpenAI and other companies began to face as they tried to scale their LLMs. From an ethics point of view many academics as well as people in the industry itself began to question the wisdom of building AI/AGI in an unregulated way. Comparisons were drawn with the development of the atom bomb during World War II and the work done by Oppenheimer and his team on the Manhattan Project. Where was the governance and regulation that was developed alongside nuclear weapons which, despite a few close shaves, have prevented nuclear Armageddon? Companies were building ethics teams to try and develop such governance models but there was often an uneasy relationship between the leadership teams whose focus was on profit and the need for an ethical approach for development. The need for ethical leadership is no more apparent when it comes to one of the activities few of us think about when using LLMs like ChatGPT and that is how these models ‘learn’. It turns out they are trained by people. But these are not the well paid engineers who live and work in San Francisco but are from third world countries like Kenya and Venezuela where labour practices are unregulated and often exploitative. As part of her research for the book Hao travels to several of these places and interviews some of the people who work long hours annotating data they are sent by the tech giants (usually via a third-party companies) describing what they see. This is not only boring and poorly paid (often just a few pennies per task) but in some cases can be hugely distressful as workers can be presented with text and images showing some of the worst forms of child sexual abuse, extreme violence, hate speech and self-harm. It’s very easy to forget, overlook or not understand that for LLMs like CharGPT to present acceptable content for our sensibilities someone, somewhere has had to filter out some of the worst forms of human degradation.

As LLMs are scaled there is a need to build more and more, ever larger data centres to house the processors that crunch the massive amounts of data, not just during training but also in their operation. Many of these large data centres are also being constructed in third-world countries where it is relatively easy to get planning permission and access to natural resources, like water for cooling, but often to the detriment of local people. In Part III, Hao discusses these aspects in detail. As new and improved versions of ChatGPT and it’s picture generation equivalent DALL-E were released and OpenAI became ever closer to Microsoft who were providing the necessary cloud infrastructure that hosted ChatGPT and DALL-E the need for ‘hyperscale’ data centres became ever greater The four largest hyperscalers – Google, Microsoft, Amazon and Meta – are now building so called ‘megacampuses’ with vast buildings containing racks of GPUs each of which will soon require 1,000 to 2,000 megawatts of power – the equivalent energy requirement of up to three and a half San Francisco’s. Such power hungry megacampuses mean that these companies can no longer meet their carbon emission targets (Google’s carbon emissions have soared by 51% since 2019 as they have invested in more and more artificial intelligence).

As Altman’s fame and influence grew his personal life inevitably began to get more attention. In September 2023 a feature writer at New York Magazine, Elizabeth Weil, published a profile of Altman which, for the first time in mainstream media, discussed his estrangement from his sister, Annie, and how financial, physical and mental health issues had caused her to turn to sex work. The New York magazine profile set side-by-side Annie’s life of financial problems with Altman’s lifestyle of expensive homes and luxury cars. Hao draws comparisons with how OpenAI (and other AI companies) ignore the anger of the data workers who try to challenge their domination by fighting for fair working conditions with how Altman seems able to do the same in ignoring his sisters cries for help. It would seem that Altman’s personal and professional lives were beginning to conspire against his so far meteoric success. In the final part of the book we see how a particular aspect of his personality lead to the events of that fateful weekend in November of 2023.

From the outside, much of what began to ensue at OpenAI after ChatGPT had propelled the company to a valuation in excess of $100B could be seen to be problems that face any company that had grown so quickly. As the spotlight on Altman himself had become ever more intense however Altman’s behaviour began to deteriorate. Often exhausted he was cracking under pressure of mounting competition as well as the punishing travel schedule he had set himself to promote OpenAI. According to Hao this pressure was causing Altman to exhibit destructive behaviour. “He was doing what he’d always done, agreeing with everyone to their face, and now, with increasing frequency, badmouthing them behind their backs. It was creating greater confusion and conflict across the company than ever before, with team leads mimicking his bad form and pitting their reports against each other“.

This, together with concerns about Altman forcing his developers to deliver new iterations of ChatGPT without sufficient testing finally drove the board, on Saturday 11th November 2023, to come to their momentous decision – “they would remove Altman and install Murati as interim CEO“. Mira Murati was OpenAIs CTO but in that role had found herself “frequently cleaning up his [Altman’s] messes“.

And so the book returns to where it started with the events of the 17th – 21st November. As we know, Altman survived what is now referred to internally as “The Blip” but pressure on him continues to mount from several directions – multiple lawsuits (including from Altman’s co-founder Elon Musk), investigations from regulators after the board investigation had observed Altman was “”not consistently candid in his communications” and increased competition, even from Microsoft who had decided to diversify its AI portfolio not wishing to put all of its AI eggs in OpenAI’s basket.

As followers of OpenAI will know, Altman and his team have gone on to deliver regular updates to ChatGPT as well as the API which can be used by developers to access its functionality. The current version (at the time of writing this review) of ChatGPT (o3-pro) is ‘multi-modal’ in that it can search the web, analyse files, reason about visual inputs, use Python, personalise responses using memory, and a whole load more. Its competitors too are releasing ever more powerful models though none (yet) claim to have achieved the holy grail of AGI. Empire of AI has captured a relatively small slice in time of the race to AGI and no doubt many more books will be written which chart the twists and turns of that race.

Empire of AI is a BIG book (nearly 500 pages with notes and index) and is the result of over 300 interviews plus a “trove of correspondence and documents” gathered by Karen Hao since she began covering OpenAI in 2019. Like many such books, you may wonder if the editing could have been a bit sharper. Perhaps reducing the number of stories and incidents would have made its points more succinctly (and in fewer pages). Ultimately however this is an important document that describes well the personalities involved in building OpenAI and the design and delivery of its products, not to mention the absolute and total belief the founders have in these products. Like the book Careless People by Sarah Wynn-Williams – which captures the power, greed and madness at Facebook during its early years – you do not come away from reading Empire of AI with much of a sense of trust or admiration for the men (for they are nearly all men) that start and run these companies. One can only hope that the steady drip of publications that are critiquing the tech industry in general and the AI companies in particular will ultimately lead to some form of change which limits and constrains the power of the people that run the companies as well as the technology itself.

I for one am not holding my breath though.

Is an AI ‘Parky’ the first step in big techs takeover of the entertainment industry?

Peter Cripps comment , , , , ,
Composite Image Created using OpenAI, DALL-E and Adobe Photoshop

Sir Michael Parkinson, who died in 2023 [1], was a much loved UK chat show host who worked at the BBC between 1971 and 1982, again between 1998 and 2004 and finally for a further three years at ITV until 2007. During that time “Parky” interviewed the great and the good (and sometimes the not so good [2]) from film, television, music, sport, science and industry. I remember Saturday nights during his first stint at the BBC not feeling complete unless we had tuned into Parkinson to see which celebrities he was interviewing that night. I was sad to hear of his passing last year but also grateful I had lived at the time to see many of his interviews and appreciate his gentle but probing interview style.

Last week however we learnt that just because you are dead, it does not mean you cannot carry on doing your job. Mike Parkinson, son of Sir Michael, has given permission to Deep Fusion Films to create an exact replica of his late father’s voice so he can virtually host a new eight-part, “unscripted series” [3] called Virtually Parkinson. The virtual Parky will be able to interview new guests based on analyses of data obtained from the real Parkinson’s back catalogue [4].

Deep Fusion Films was founded in 2023 and makes a big play about its ethical credentials. On its website [5] it says it aims to “establish comprehensive policies that promote the legal and ethical integration of AI in production”. Backing this up, their virtual Parky will be created with the full support and involvement of Sir Michael’s family and estate. 

So far, so ethical, right and proper, however…

Only last year, concerns over the use (and potential misuse) of AI in the film industry led to a strike by actors and writers. Succession actor Brian Cox made the statement that using AI to replicate an actor’s image and use it forever is “identity theft” and should be considered “a human rights issue” [6].

Hollywood stars like Scarlett Johansson, Tom Hanks, Tom Cruise and Keanu Reeves, have already become the subject of unauthorised deepfakes and in June of this year the Internet Watch Foundation(IWF) warned that AI-generated videos of child sexual abuse could indicate a ‘stark vision of the future’ [7].

Clearly, where Deep Fusion Films are right now, i.e. producing ethically sourced and approved imitations of celebrities voices, and where AI generated porn is threatening to take us are poles apart but…

Technology always creeps into our lives like this. A small seemingly insignificant event which we find amusing and mildly distracting entertains us for a while but then suddenly, it has become the way of all things and has fallen into the hands of ‘bad actors’. At this point, there is often no going back.

Witness how Facebook started out as an innocuous site called Facemash, created by a second-year student at Harvard University called Mark Zuckerberg, that compared two student photos side-by-side to determine who was “hot” and who was “not.” Actually this was always a questionable use case in my opinion, but I guess an indication of what went down as acceptable behaviour in Ivy League universities of the early 2000s!

Today Meta (who now owns Facebook) is the seventh largest company in the world by market capitalisation worth, at the time of writing, $1.497 T [8]. Zuckerberg’s vision for Meta, outlined in a letter to shareholders this August, is that it will become a virtual reality platform that merges the physical and digital worlds forever transforming how we interact, work, and socialise [9]. Inevitably a major part of this vision is that artificial intelligence (or even, if Zuckerberg gets his way, artificial general intelligence) will be there to “enhance user experiences”.

Facebook, and now Meta, is surely the canonical example of how a small and seemingly insignificant company from the US east coast has grown in a mere 20 years to become a largely unregulated west coast tech behemoth with over three billion active monthly users [10].

If Facebook was just used for sharing pictures of cats and dogs that would be one thing but, during its short history, it has been found guilty of spreading fake news, changing voting behaviour in key elections around the world, affecting peoples mental health as well as spreading violent and misogynistic (and deepfake) videos.

It seems like we never learn. Governments and legal systems around the world never react fast enough to the pace of technological change and are always playing catchup having to mop up the tech companies misdemeanours after they have occurred rather than regulating against tech companies in the first place. Financial penalties are one thing but these pale into insignificance alongside the gargantuan profits such companies make and anyway, no amount of fines can undo the negative effects they and their leaders have on peoples lives.

So how does the rise of the tech behemoths like Facebook, Google and X presage what might happen in the creative industries and their use of technology, especially AI?

I don’t know what proportion of a Hollywood movies costs goes to actors salaries. It is obviously not the only cost or even the largest cost however with actors like Tom Cruise, Keanu Reeves and Will Smith able to command salaries for a single film in excess of $100M [11] salaries are clearly not insignificant. It must be very tempting for movie producers to be thinking why not invest a bit more in special effects and just create a whole new actor from scratch. After all, that’s precisely what Walt Disney did with Mickey Mouse who never got paid a dime.

How long is it before we cross a red line and a movies special effects goes the whole way and uses CGI to create the characters in a completely AI scripted and generated film? Huge upfront costs (for now, but these will drop) but no ongoing costs of having to pay actors for re-runs or streaming rights etc.

I don’t know how long it might take or whether we will ever get there. Maybe the technology will never be good enough (unlikely) or maybe we will wake up to what we are doing and create some sort of legal/ethical framework that prevents such things occurring (equally unlikely I fear).

We are beginning to rub up against some pretty fundamental questions not just about how we should be using AI, especially in the creative industries, but also what it actually means to be human if we let our machines overwhelm us to the extent that our creative selves are usurped by the very things that creativity has built.

This is a hugely important question which I hope to explore in future posts. 

Notes

  1. Sir Michael Parkinson obituary,https://www.theguardian.com/media/2023/aug/17/sir-michael-parkinson-obituary
  2. Michael Parkinson speaks out on Savile scandal, https://www.itv.com/news/calendar/2012-12-01/michael-parkinson-speaks-out-on-savile-scandal
  3. AI-replicated Michael Parkinson to host ‘completely unscripted’ celebrity podcast, https://news.sky.com/story/ai-replicated-michael-parkinson-to-host-completely-unscripted-celebrity-podcast-13243556
  4. Michael Parkinson is back, with an AI voice that can fool even his own family, https://www.theguardian.com/media/2024/oct/26/michael-parkinson-virtually-ai-replica-chatshow
  5. Deep Fusion Films is a dynamic production company at the forefront of television and film,https://www.deepfusionfilms.com/about
  6. Succession star Brian Cox on the use of AI to replicate actors: ‘It’s a human rights issue’,https://news.sky.com/story/succession-star-brian-cox-on-the-use-of-ai-to-replicate-actors-its-a-human-rights-issue-12999168
  7. AI-generated videos of child sexual abuse a ‘stark vision of the future’, https://www.iwf.org.uk/news-media/news/ai-generated-videos-of-child-sexual-abuse-a-stark-vision-of-the-future/
  8. Largest Companies by Marketcap,https://companiesmarketcap.com
  9. Mark Zuckerberg’s Letter: Meta’s Vision Unveiled,https://medium.com/@ahmedofficial588/mark-zuckerbergs-letter-meta-s-vision-unveiled-2b48a57a2743
  10. Facebook User & Growth Statistics,https://backlinko.com/facebook-users
  11. 20 Highest Paid Actors For a Single Film,https://thecinemaholic.com/highest-paid-actors-for-a-single-film/

Tech: The Missing Generation

Peter Cripps comment , , , , , , , , , , ,

I’ve recently been spending a fair bit of time in hospital. Not, thankfully, for myself but with my mother who fell and broke her arm a few weeks back which has resulted in lots of visits to our local Accident & Emergency (A&E)  department as well as a short stay in hospital whilst they pinned her arm back in place.

nhs hospital
An elderly gentleman walks past an NHS hospital sign in London. Photograph: Cate Gillon/Getty Images

Anyone who knows anything about the UK also knows how much we value our National Health Service (NHS). So much so that when it was our turn to run the Olympic Games back in 2012 Danny Boyle’s magnificent opening ceremony dedicated a whole segment to this wonderful institution featuring doctors, nurses and patients dancing around beds to music from Mike Oldfield’s Tubular Bells.

nhs london 2012 olympics
Olympic Opening Ceremony NHS Segment – Picture Courtesy the International Business Times

The NHS was created out of the ideal that good healthcare should be available to all, regardless of wealth. When it was launched by the then minister of health, Aneurin Bevan, on July 5 1948, it was based on three core principles:

  • that it meet the needs of everyone
  • that it be free at the point of delivery
  • that it be based on clinical need, not ability to pay

These three principles have guided the development of the NHS over more than 60 years, remain at its core and are embodied in its constitution.

nhs constitution
NHS Constitution Logo

All of this, of course, costs:

  • NHS net expenditure (resource plus capital, minus depreciation) has increased from £64.173 billion in 2003/04 to £113.300bn in 2014/15. Planned expenditure for 2015/16 is £116.574bn.
  • Health expenditure (medical services, health research, central and other health services) per capita in England has risen from £1,841 in 2009/10 to £1,994 in 2013/14.
  • The NHS net deficit for the 2014/15 financial year was £471 million (£372m underspend by commissioners and a £843m deficit for trusts and foundation trusts).
  • Current expenditure per capita for the UK was $3,235 in 2013. This can be compared to $8,713 in the USA, $5,131 in the Netherlands, $4,819 in Germany, $4,553 in Denmark, $4,351 in Canada, $4,124 in France and $3,077 in Italy.

The NHS also happens to be the largest employer in the UK. In 2014 the NHS employed 150,273 doctors, 377,191 qualified nursing staff, 155,960 qualified scientific, therapeutic and technical staff and 37,078 managers.

So does it work?

From my recent experience I can honestly say yes. Whilst it may not be the most efficient service in the world the doctors and nurses managed to fix my mothers arm and hopefully set her on the road to recovery. There have been, and I’m sure there will be more, setbacks but given her age (she is 90) they have done an amazing job.

Whilst sitting in those A&E departments whiling away the hours (I did say they could be more efficient) I had plenty of time to observe and think. By its very nature the health service is hugely people intensive. Whilst there is an amazing array of machines beeping and chirping away most activities require people and people cost money.

The UK’s health service, like that of nearly all Western countries, is under a huge amount of pressure:

  • The UK population is projected to increase from an estimated 63.7 million in mid-2012 to 67.13 million by 2020 and 71.04 million by 2030.
  • The UK population is expected to continue ageing, with the average age rising from 39.7 in 2012 to 42.8 by 2037.
  • The number of people aged 65 and over is projected to increase from 10.84m in 2012 to 17.79m by 2037. The number of over-85s is estimated to more than double from 1.44 million in 2012 to 3.64 million by 2037.
  • The number of people of State Pension Age (SPA) in the UK exceeded the number of children for the first time in 2007 and by 2012 the disparity had reached 0.5 million (though this is projected to reverse by).
  • There are an estimated 3.2 million people with diabetes in the UK (2013). This is predicted to reach 4 million by 2025.
  • In England the proportion of men classified as obese increased from 13.2 per cent in 1993 to 26.0 per cent in 2013 (peak of 26.2 in 2010), and from 16.4 per cent to 23.8 per cent for women over the same timescale (peak of 26.1 in 2010).

The doctors and nurses that looked after my mum so well are going to be coming under a increasing pressures as this ageing and less healthy population begins to suck ever more resources out of an already stretched system. So why, given the passion everyone has about the NHS, isn’t there more of a focus on getting technology to ease the burden of these overworked healthcare providers?

Part of the problem of course is that historically the tech industry hasn’t exactly covered itself with glory when it comes to delivering technology to the healthcare sector (I’m thinking the NHS National Programme for IT and the US HealthCare.gov system as being two high profile examples). Whilst some of this may be due to the blunders of government much of it is down to a combination of factors caused by both the providers and consumers of healthcare IT mis-communication and not understanding the real requirements that such complex systems tend to have.

In her essay How to build the Next Unicorn in Healthcare the entrepreneur Yasi Baiani   sets out six tactical tips for how to build a unicorn* digital startup. In summary these are:

  1. Understand the current system.
  2. Know your customers.
  3. Have product hooks.
  4. Have a clear monetization strategy and understand your customers’ willingness-to-pay.
  5. Know the rules and regulations.
  6. Figure out what your unfair competitive advantage is.

Of course, these are strategies that actually apply to any industry when trying to bring about innovation and disruption – they are not unique to healthcare. I would say that when it comes to the healthcare industry the reason why there has been no Uber is because the tech industry is ignoring the generation that is in most need of benefiting from technology, namely the post 65 age group. This is the age group that struggle most with technology either because they are more likely to be digitally disadvantaged or because they simply find it too difficult to get to grips with it.

As the former Yahoo chief technology officer Ashfaq Munshi, who has become interested in ageing tech says:

“Venture capitalists are too busy investing in Uber and things that get virality. The reality is that selling to older people is harder, and if venture capitalists detect resistance, they don’t invest.”

Matters are not helped by the fact that most tech entrepreneurs are between the ages of 20 and 35 and have different interests in life than the problems faced by the aged. As this article by Kevin Maney in the Independent points out:

“Entrepreneurs are told that the best way to start a company is to solve a problem they understand. It makes sense that those problems range from how to get booze delivered 24/7 to how to build a cloud-based enterprise human resources system – the tangible problems in the life and work of a 25- or 30-year-old.”

If it really is the case that entrepreneurs only look at problems they understand or are on their immediate event horizon then clearly we need more entrepreneurs of my age group (let’s just say 45+). We are the people either with elderly parents, like my mum, who are facing the very real problems of old age and poor health and who themselves will very soon be facing the same issues.

A recent Institute of Business Value report from IBM makes the following observation:

“For healthcare in particular, the timing for a game changer couldn’t be better. The industry is coping with upheaval triggered by varied economic, societal and industry influences. Empowered consumers living in an increasingly digital world are demanding more from an industry that is facing growing regulation, soaring costs and a shortage of skilled resources.”

Rather than fearing the new generation of cognitive systems we need to be embracing them and ruthlessly exploiting them to provide solutions that will ease all of our journeys into an ever increasing old age.

At  SXSW, which is running this week in Austin, Texas IBM is providing an exclusive look at its cognitive technology called Watson and showcasing a number of inspiring as well as entertaining applications of this technology. In particular on Tuesday 15th March there is a session called Ageing Populations & The Internet of Caring Things  where you can take a look at accessible technology and how it will create a positive impact on an aging person’s quality of life.

Also at SXSW this year President Obama gave a keynote interview where he called for action in the tech world, especially for applications to improve government IT. The President urged the tech industry to solve some of the nation’s biggest problems by working in conjunction with the government. “It’s not enough to focus on the cool, next big thing,” Obama said, “It’s harnessing the cool, next big thing to help people in this country.”

obama-sxsw
President Barack Obama speaks during the 2016 SXSW Festival at Long Center in Austin, Texas, March 11, 2016. PHOTO: NEILSON BARNARD/GETTY IMAGES FOR SXSW

It is my hope that with the vision that people such as Obama have given the experience of getting old will be radically different 10 or 20 years from now and that cognitive and IoT technology will make all of out lives not only longer but more more pleasant.

* Unicorns are referred to companies whose valuation has exceeded $1 billion dollars.