I have lived through an interesting technological era. When I was young, many families still used black and white televisions, phones were attached to walls, and the internet hadn’t been created. Much has changed in the intervening years, and now we’re living in a time when technology that didn’t exist during my youth is deeply embedded in our lives. As Michio Kaku put it, “Today, your cell phone has more computer power than all of NASA back in 1969, when it placed two astronauts on the moon.”1
Over the last few years, something new entered the frame as ‘AI’ gave things a jolt. Large Language Models (LLMs) like ChatGPT spit out text at alarming rates, while text-to-image generators hurled pics hither and yon. Their outputs were all derivatives, but the speed with which they were produced was unnerving.
While the history of AI traces back to the days of Alan Turing, it was a relatively quiet corner of technological development until recently. As such, we’re just starting to get familiar with these offerings and to consider repercussions and ethical concerns.
AI’s supporters offer a laundry list of potential economic gains, but I’m a bit of a sceptic. I like to think of it in terms of the Luddites, in that I’d like to know where the gains might accrue before getting excited about any new tech. Let’s come back to this.
This survey suggests a consensus from business leaders that AI will consume more jobs than it creates over the next few years.
ECONOMIC POSSIBILITIES
An April 2023 report from Goldman Sachs estimated generative AI could lift global GDP by 7% (nearly USD 7 trillion) over the next decade while increasing productivity by 1.5%. The same report hailed AI’s “ability to generate content that is indistinguishable from human-created output.” (emphasis mine) How that claim lands with you might tell you what to make of their projections.2
McKinsey, the global consulting firm that’s paid nearly USD 1 billion to settle claims for their part in the US opioid crisis, runs an annual survey of business leaders around the evolution of AI use. An important concern the survey addresses is AI’s potential impact on jobs. Their latest survey, published in August 2023, asked respondents whether they expected AI to lead to changes in their employee headcounts over the next three years, and if so, to what degree. While 15% expected their headcount to increase by 3% or more, 43% expected to reduce their numbers by at least that much. This survey suggests a consensus from business leaders that AI will consume more jobs than it creates over the next few years.
The same survey asked respondents what generative AI-related risks were most relevant. Over half agreed that inaccuracy and cybersecurity are top concerns. These were closely followed by intellectual property infringement, regulatory compliance, explainability, and personal/individual privacy. All these risks speak to AI-related trust concerns.
Another important question from the survey was about the need for “reskilling”— retraining employees to meet shifting needs due to the use of AI. The findings here were shocking. Nearly three-fourths of respondents expect that at least 6% of employees will need reskilling, while over half of that group expects to reskill at least 20% of their workforce. My initial reaction here is that we all might want to back away from the AI cliff and take a moment to calm down and think about what we’re saying and doing. But while these figures feel inflated to me, they come from people who make these decisions. Anyway, the sort of upheaval these figures suggest is dizzying. If this is where we are headed, human resources departments had better be ready. More importantly, I hope workers are considering the risks AI may be fostering around their employment, and that they’re thinking about what they might do next if they get tagged for reskilling, as I fear many of those people will not land softly. In fact, I suspect ‘reskilling’ is largely a spillover category to make internal projections of headcount reductions look less daunting. We’ll see. 3
Biggest ever global strike to make Amazon Pay
In November 2023, Amazon workers globally staged a Black Friday strike, demanding better pay and working conditions. Part of the “Make Amazon Pay” campaign, the protest highlighted concerns over automation, job security, and corporate accountability, urging Amazon to address the negative impacts on workers and communities.
Photo: Zeynep Demir Aslim / Alamy Stock Photo
A finding from edX’s 2023 AI Survey might put the reskilling projections in context. It found that 49% of CEOs believed “most” or “all” of their work should be automated or replaced by AI. That revelation reminded me of a scene from Office Space, a comedy in which a downsizing consultant asks a manager who seems to do nothing, “What would you say you do here?”4
There is always a tension between who controls technology and whether the gains are widely shared.
In September 2023, MIT Economist Daron Acemoglu argued that we should use AI to supplement workers by making machines “more and more useful to humans,” but whether they’ll see gains from such efforts depends on whether and how any economic benefits will be shared. He noted that we can take these technologies in many different directions, so it’s possible they could create broad-based benefits, but they could also “enrich and empower a very narrow elite.” Acemoglu says that, so far, it’s primarily been used to replace workers and increase community surveillance.
Acemoglu shared a history lesson to put the moment in perspective, noting that gains from the Industrial Revolution weren’t widely shared for a century. Labour conditions initially worsened as working hours increased by 20%. As he put it, we are “immeasurably more prosperous, healthier, and more comfortable today than people were 300 years ago,” before adding that nothing assured those shared gains. The economist closed his session noting that we face similar choices to those in the past with the potential for serious consequences and that there is always a tension between who controls technology and whether the gains are widely shared.5
More recently, Acemoglu drafted a research paper for the National Bureau of Economic Research on the macroeconomics of AI. His study estimates modest productivity gains over the next decade as AI systems improve the productivity of low-skill workers, and the potential for increased inequality. He further finds that AI is unlikely to worsen income inequality, thanks to a relatively equal distribution across demographic groups, but that AI will likely “widen the gap between capital and labour income.”6 7
The IMF published a study in January 2024 estimating that we could replace or “complement” 40% of the global labour force with AI. While historical technological shifts have tended to affect lower-skilled workers, AI is expected to hit other workers more. This could put greater risks on advanced economies, but they are also expected to have more opportunities to leverage these technologies. That said, it’s important to note that workers are not the same as economies, so the ability of economies to leverage technologies may not matter much for displaced workers. On the upside, this study seemed to suggest far fewer job losses than the survey from McKinsey.8
Digital money or digital control?
The Atlantic Council reports that three countries have fully launched a central bank digital currency (CBDC), with 36 pilots underway, including the digital euro, and 134 in the exploration phase. While CBDCs promise to improve financial inclusion and streamline payments, they raise concerns about transaction tracking, state surveillance, financial control, and cybersecurity.
Image: dem10 / iStock
Workers are not the same as economies, so the ability of economies to leverage technologies may not matter much for displaced workers.
In May 2024, two trusted writers questioned whether AI had been overhyped. Julia Angwin said the “question isn’t really whether AI is too smart and will take over the world. It’s whether AI is too stupid and unreliable to be useful.” Christopher Mims suggested AI might fall well short of expectations and bring significant disappointment in terms of its capabilities and financial returns. On June 1, the New York Times reported that Google had rolled back access to their AI search overview after it “produced a litany of untruths and errors — including recommending glue as part of a pizza recipe and suggesting that people ingest rocks for nutrients.”9 10 11
ENVIRONMENTAL CONCERNS
A big question for AI’s proponents surrounds its resource use. Some believe that the increasing capabilities of these platforms will help us reduce energy consumption. As OpenAI’s leader, Sam Altman, put it, “If we have to spend even 1% of the world’s electricity training powerful AI, and that AI helps us figure out how to get to non-carbonbased energy or to do carbon capture better, that would be a massive win.” That might sound nice, but It’s not something that OpenAI is pursuing. It’s also not something Altman is even advocating for others to do. Instead, it was a hypothetical that was shot from the hip, seemingly to mollify AI’s supporters. If anyone did try that ‘moonshot,’ it would be an expensive, risky lottery ticket.12
Back in January 2020, Microsoft’s leadership made a bold promise, the company would be carbonnegative by 2030. That announcement followed a 2019 investment of USD 1 billion in OpenAI. In November 2022, OpenAI launched ChatGPT, and with it, the ongoing AI madness in which obscene wealth transfers and absurd energy consumption became the norm. Microsoft welcomed both these matters by upping its investment in OpenAI to USD 13 billion in 2023, as its CO2 emissions increased nearly 30% (from the 2020 baseline).
With six years left to hit their target, Microsoft inked a deal to support the development of 10.5 gigawatts of new renewable energy. The USD 11.5 to USD 17 billion investment will create half as much renewable energy as the entire state of California had in 2022. Unfortunately, roughly a quarter of that energy will go to things like AIgenerated meeting notes, rather than reducing existing emissions to help combat climate change.13 14 15 16 17
As Google and Microsoft shift to AI-based search responses, the related energy consumption and emissions are growing. AI researcher Sasha Luccioni says that AI uses “orders of magnitude more energy” than traditional search engines. Relatedly, data centres currently use roughly 1.5% of global energy, but that figure is projected to double by 2026. By then, generative AI is expected to use ten times the energy it used in 2023.18
How much electricity does AI consume?
The International Energy Agency's 2024 forecast estimates that the energy demand from data centres, cryptocurrency, and AI, which constituted nearly 2% of global energy demand in 2022, could double by 2026, possibly matching Japan's total electricity consumption. Generative AI, among the fastest-growing energy consumers, requires as much electricity annually to train models like GPT-3 as about 130 US homes use in a year.
Photo: Tetra Images / Alamy Stock Photo
There is deep pressure from companies — that are basically promising God and delivering email prompts — to make some return on investment in this technology.
THE PROBLEM OF TRUST
One of the biggest challenges I see for AI is the matter of trust. After spending multiple years taking data from the internet, OpenAI’s leadership has recently started inking deals with several newspapers. Make of that what you will.19
Sam Altman led the shift of OpenAI from a traditional non-profit to a hybrid form that’s now valued at USD 80 billion and has reportedly leveraged the firm’s success to grow a personal fortune. Now, a growing contingent of the firms he’s invested in are doing business with OpenAI, giving him multiple potential conflicts of interest.20 21
There’s also the matter of Altman’s brief firing in November 2023. A former board member, Helen Toner, claims that he was let go for repeatedly lying to the board. Since Altman’s return, OpenAI has experienced an “alarming” number of exits. This included multiple leaders from a safety and security committee that disbanded amid complaints of never receiving access to promised resources. As Altman is the de facto voice of AI, his words and actions have inflated weight. It’s worth questioning where AI will go if he continues to light the way.22 23
OpenAI’s major partner, Microsoft, has also had its share of worrisome AI-related headlines. One major concern comes from Recall, an AI-enabled feature coming to Windows 11. Microsoft’s CEO called it a “photographic memory” for your computer, as it saves a picture of your screen every five seconds. However, cyber security experts have called it a hacker’s dream come true. Multiple ethical hackers have proven serious issues with the system’s security, including one who found a way to remotely access the pictures without having admin privileges. In the days following that revelation, Microsoft announced the feature would be turned off by default.24 25 26
Crushing creativity
Apple’s 2024 iPad Pro ad featured a powerful hydraulic press methodically crushing vibrant objects that symbolise creativity: a grand piano, a trumpet, a sculpture, books and paint palettes, leaving only the sleek tablet unscathed. The ad sparked a storm of online backlash, with actor Hugh Grant calling it “the destruction of human experience” on social media. The ad exposed Apple’s disconnect with the public’s anxieties about technology and AI taking over humanity.
Source: Apple
It is also worth considering where AI’s funding comes from. In a recent talk, Signal’s president, Meredith Whitaker, shared concerns about privacy due to “the cold, hard business model of tech,” before adding, “If we were a for-profit it is very likely that we would be pushed to erode privacy guarantees in an industry where collecting, selling, and making use of personal data is the primary economic driver.” Whitaker further noted that AI models cost hundreds of millions of dollars to set up, “So there is deep pressure from companies — that are basically promising God and delivering email prompts — to make some return on investment in this technology.”27
It is easy to guess what the Luddites would think of investors solely focused on market returns. I doubt any products such efforts create will benefit humanity or the planet. As David Karpf put it, “We don’t have to put our faith in the techno-futurists. In fact, based on their track records, we ought to regard them with deep suspicion.”28
I do believe we could use these systems to produce social and environmental gains, but we would have to be far more thoughtful and intentional with their use.
I do believe we could use these systems to produce social and environmental gains, but we would have to be far more thoughtful and intentional with their use as our current approach is burning up resources with little to show for it. They could probably help us with things like assessing our collective energy use and finding meaningful opportunities for reduction. I am sure they could help us improve the outcomes of our global food systems that currently waste enough food to feed roughly two billion people, while around threequarters of a billion go hungry (lost and wasted food accounts for 38% of the global food system’s energy consumption).29 30
Pursuing such changes might require us to start thinking collectively across boundaries where collaboration isn’t the norm, like between competitor firms or across nations. That might sound like a non-starter, but given the resources used by these systems, and the many other problems around AI, continuing to allow people to use them for personal enjoyment, or worse, is a road to ruin. ∞
CHRIS OESTEREICH
Chris Oestereich is a writer, lecturer, and circular economy leader. He is the publisher of the Wicked Problems Collaborative, and a lecturer at Thammasat University’s School of Global Studies, where he teaches sustainability and social innovation. He’s also a co-founder of the Systemic Design Studio and the Circular Design Lab, efforts that work to bring systemic design to communities. Chris formerly led zero waste programmes in the US grocery industry and he now helps organisations shift towards the circular economy via his firm, Linear to Circular. His new project is Morph, a social enterprise that upcycles waste into useful, everyday goods.
JULY 2024 | ISSUE 12
NAVIGATING THE AI TERRAIN
- Kaku, Michio. Physics of the Future: The Inventions That Will Transform Our Lives. Penguin UK, 2012.
- “Generative AI Could Raise Global GDP by 7%.” Goldman Sachs, 5 Apr 2023. https://www.goldmansachs.com/intelligence/pages/ generative-ai-could-raise-global-gdp-by-7-percent.html.
- Hanselman, Heather. “The State of AI in 2023: Generative AI's Breakout Year.” McKinsey & Company, 1 Aug 2023. https://www. mckinsey.com/capabilities/quantumblack/our-insights/the-state-ofai- in-2023-generative-ais-breakout-year.
- “edX Survey Finds Nearly Half (49%) of CEOs Believe Most or All of Their Role Should be Automated or Replaced by AI.” edX, 19 Sept 2023. https://press.edx.org/edx-survey-finds-nearly-half-49-ofceos- believe-most-or-all-of-their-role-should-be-automated-orreplaced- by-ai.
- Dizikes, Peter. “Who Will Benefit From AI?” MIT News, 29 Sept 2023. https://news.mit.edu/2023/who-will-benefit-ai-machineusefulness- 0929.
- “Daron Acemoglu is not having all this AI hype.” Financial Times, 28 May 2024. https://www.ft.com/content/b375115f-278f-43a3-9a26- 31d75e5cd319.
- Acemoglu, Daron. “The Simple Macroeconomics of AI.” National Bureau of Economic Research, May 2024. https://www.nber.org/ papers/w32487.
- Georgieva, Kristalina. “AI Will Transform the Global Economy. Let’s Make Sure It Benefits Humanity.” International Monetary Fund, 14 Jan 2024. https://www.imf.org/en/Blogs/Articles/2024/01/14/ ai-will-transform-the-global-economy-lets-make-sure-it-benefitshumanity.
- Angwin, Julia. “Press Pause on the Silicon Valley Hype Machine.” The New York Times, 15 May 2024. https://www.nytimes. com/2024/05/15/opinion/artificial-intelligence-ai-openai-chatgptoverrated- hype.html.
- Mims, Christopher. “The AI Revolution Is Already Losing Steam.” The Wall Street Journal, 31 May 2024. https://www.wsj.com/tech/ai/ the-ai-revolution-is-already-losing-steam-a93478b1.
- Grant, Nico. “Google Rolls Back AI Search Feature After Flubs and Flaws.” The New York Times, 1 Jun 2024. https://www.nytimes. com/2024/06/01/technology/google-ai-overviews-rollback.html.
- Eastwood, Brian. “Sam Altman Believes AI Will Change the World (And Everything Else).” MIT Sloan School of Management, 8 May 2024. https://mitsloan.mit.edu/ideas-made-to-matter/sam-altmanbelieves- ai-will-change-world-and-everything-else.
- Smith, Brad. “Microsoft will be carbon negative by 2030.” Microsoft, 16 Jan 2020. https://blogs.microsoft.com/blog/2020/01/16/ microsoft-will-be-carbon-negative-by-2030/.
- Robinson, Dan. “Microsoft’s Carbon Emissions up Nearly 30% Thanks to AI.” The Register, 16 May 2024. https://www.theregister. com/2024/05/16/microsoft_co2_emissions/.
- O’Brien, Matt, et al. “AI Tools Fueled a 34% Spike in Microsoft’s Water Consumption, and One City With Its Data Centers Is Concerned About the Effect on Residential Supply.” Fortune, 9 Sept 2023. https://fortune.com/2023/09/09/ai-chatgpt-usage-fuelsspike- in-microsoft-water-consumption/.
- Calma, Justine. “Microsoft Made the Biggest Renewable Energy Agreement Ever to Fuel Its AI Ambitions.” The Verge, 3 May 2024. https://www.theverge.com/2024/5/2/24147153/microsoft-ai-datacenter- record-renewable-energy-purchase.
- Novet, Jordan. “Microsoft’s $13 Billion Bet on OpenAI Carries Huge Potential Along With Plenty of Uncertainty.” CNBC, 9 Apr 2023. https://www.cnbc.com/2023/04/08/microsofts-complex-bet-onopenai- brings-potential-and-uncertainty.html.
- Parshall, Allison. “What Do Google’s AI Answers Cost the Environment?” Scientific American, 11 Jun 2024. https://www. scientificamerican.com/article/what-do-googles-ai-answers-costthe- environment/.
- Koenig, Melissa. “OpenAI Is Making Rapid-Fire Media Deals: Every News Publication on ChatGPT Now.” Observer, 31 May 2024. https://observer.com/2024/05/openai-news-media-deal-list/.
- “Our Structure.” OpenAI. https://openai.com/our-structure/. Accessed 13 Jun 2024.
- Metz, Cade and Tripp Mickle. “OpenAI Completes Deal That Values the Company at $80 Billion.” The New York Times, 16 Feb 2024. https://www.nytimes.com/2024/02/16/technology/openaiartificial- intelligence-deal-valuation.html.
- Morris, Chris. “OpenAI Resignations Are Reaching an Alarming Level. Here Are 11 Key People Who Have Left.” Fast Company, 17 May 2024. https://www.fastcompany.com/91126785/openairesignations- are-reaching-an-alarming-level-here-are-11-keypeople- who-have-left.
- Tremayne-Pengelly, Alexandra. “Inside OpenAI’s 9-Person Safety Committee Led by All-Powerful Sam Altman.” Observer, 29 May 2024. https://observer.com/2024/05/inside-openais-9-personsafety- committee-led-by-all-powerful-sam-altman/.
- Burgess, Matt. “This Hacker Tool Extracts All the Data Collected by Windows’ New Recall AI.” Wired, 4 Jun 2024. https://www.wired. com/story/total-recall-windows-recall-ai/.
- Greenburg, Andy. “Microsoft Will Switch Off Recall by Default After Security Backlash.” Wired, 7 Jun 2024. https://www.wired.com/ story/microsoft-recall-off-default-security-concerns/.
- Greenburg, Andy. “Microsoft’s Recall Feature Is Even More Hackable Than You Thought.” Wired, 6 Jun 2024. https://www.wired.com/ story/microsoft-windows-recall-privilege-escalation/.
- Morrone, Megan. “Signal’s Meredith Whittaker: AI Is a Privacy Nightmare.” Axios, 5 Jun 2024. https://www.axios. com/2024/06/05/artificial-intelligence-meredith-whittaker-signaltelegram.
- Karpf, Dave. “Bullet Points: Internet Time Ain’t What It Used to Be.” Dave Karpf, 23 May 2024. https://davekarpf.substack.com/p/bulletpoints- internet-time-aint.
- “5 Facts About Food Waste and Hunger.” World Food Programme, 2 Jun 2020. https://www.wfp.org/stories/5-facts-about-food-wasteand- hunger.
- “International Day of Awareness on Food Loss and Waste Reduction 29 September.” United Nations. https://www.un.org/en/ observances/end-food-waste-day#:~:text=the%20consumption%20 level.-,An%20estimated%2017%20percent%20of%20total%20 global%20food%20production%20is,in%20the%20global%20 food%20system. Accessed 13 Jun 2024.
