Eighteen months ago I went to a party in San Francisco that was thrown to celebrate generative AI as the next industrial revolution. The mood was cheerfully nihilistic. AI was about to demolish our way of life, said one partygoer. We were like farmers tending to our crops, unaware of the machinery that was on its way to chew us all up.
Safe to say, generative AI hasn’t chewed up much of anything yet. Accountants, designers, software engineers, film-makers, interpreters and all the other professions told to expect carnage are still in employment. Elections have not been stymied. The world is still turning. Those early warnings are starting to sound like a weird form of marketing.
Silicon Valley tends to be associated with optimism. The indefatigable sense that the world is on an upward trajectory is one of the tech sector’s more loveable qualities. When starry-eyed plans don’t pan out – Elon Musk’s claim that manned crafts would be flying to Mars by 2024, say – the world can extend grace. There is an understanding that optimistic ambition is a good thing.
But optimism is not the only mindset that California breeds. Across the tech sector, there is also a contingent who are driven by fear.
At the sharpest end of this are the survivalists – those who worry about the collapse of society. For some, this means buying up land in New Zealand or stockpiling water.
For others, it can be a business strategy. Software/consultancy company Palantir is known for using quarterly earnings to tell investors about the potential for global destruction. Existential musings add to its allure. Palantir is still described as “mysterious” despite being a public company and more than two decades old.
Fear mongering directed at tech products is not necessarily unhelpful. Calling social media addictive and privacy invading might worry users but it doesn’t put off advertisers.
[ To make artificial intelligence safer, use it moreOpens in new window ]
Look at Facebook. Its share price fell in 2018 after the revelation that Cambridge Analytica was harvesting user data and using it to conduct experiments that supposedly altered electoral outcomes. Not only did the share price recover in the space of a year, the company now trades with a market value twice as high.
Being regarded as sufficiently powerful to affect global politics made the platform sound more impressive – even if it wasn’t true. (There is still little evidence that “psychographic” data gathering swayed voters).
In AI, worriers have found something on which they can pin all their fears. Last year, OpenAI’s own co-founder, Sam Altman, joined a group of scientists and other executives in signing a letter stating that the risk of extinction from AI should become a global priority.
Other tech leaders called for research to be put on pause for six months due to “profound risks to society and humanity”. Goldman Sachs declared that 300 million full-time jobs could be automated away by the technology.
Much of this distress is no doubt genuine. But it has the side effect of priming us to be awed and then disappointed by the technology.
When OpenAI released Sora, which can generate AI videos, it was described by a reviewer as “one step closer to the end of reality itself”. Never mind that a film-maker who has used it found it less impressive.
As with any marketing, bombastic claims tend to fall apart once people try things for themselves. As generative AI is put into more of our hands – via gadgets or Google Docs or multimedia platforms – the questions about whether this is all hype are kicking up a notch.
Some of the earliest consumer products available, such as Humane’s $699 AI clip-on pin, are proving unpopular. Tech news website, the Verge reports that in the past three months, more Humane pins were returned than sold.
Meta’s Ray-Ban AI sunglasses have received better press. The glasses can tell you what you are looking at by taking a photo and identifying the item in it.
But this function, while impressive, is not perfect. When I tried a pair I found the earphone speaker function more useful. The rest of the San Francisco bureau seemed to feel the same way – trying the glasses on, dutifully using them to identity what they were looking at and then handing them back to me.
One day, perhaps, the glasses will translate street signs, offer directions and help those with impaired vision. But commercial applications of new technology are not instant.
We are in the early phase where ideas are still being tested out. The difficulty is squaring this with the message that the technology is already terrifying. We might all have more patience waiting for AI’s killer app if we hadn’t been told repeatedly it could kill us all. – Copyright The Financial Times Limited 2024
