The AI Clones Are Coming! But It’s Not All Bad

(Bloomberg Opinion) — Mark Sewards, a British member of parliament in the northern city of Leeds, recently launched what he called the country’s first AI prototype of an MP. Cue the backlash. X users were the most vitriolic, calling Sewards “lazy” and branding the project “appalling.” The press sneered at his “weird” and “dithering” chatbot, complaining it couldn’t follow a Leeds accent.

Then again, most politicians dodge questions with robotic answers, and many wouldn’t make out the heavy twang of Northern England. “Me neighbour’s lad’s blocked’t ginnel at’t back wi an old settee and he won’t do owt about it,” a Guardian journalist asked Sewards’ bot at one point. The AI suggested they call the police to report an abandoned vehicle, which was a better answer than I would have given.

The 35-year-old Labour MP is a pioneer in using AI to squeeze more out of their job. More people are using artificial intelligence to create digital replicas of themselves and conduct their work at a greater scale. Right-wing influencer Dave Rubin even has an AI clone presenting his YouTube show while he’s on vacation this month. There are obvious costs to this, from the erosion of authenticity to lower quality services. But don’t also forget the benefits to those with limited resources where chatbots offer a new way of engaging with others, however dystopian that feels to us humans.

I’ll admit to having misgivings about the side effects that generative AI is already bringing to bear, from an erosion in critical thinking to toxic attachments. But Sewards’ spokesman says his new chatbot is an “addition” to the ways people can communicate with him, like email and voicemail, and that seems both a sensible and useful way to capitalize on artificial intelligence.

For people who make money off their expertise, clones offer a route to scaling up their business. Brodie Sharpe is a physiotherapist in Melbourne, Australia, with an entrepreneurial streak. Having fallen prey to a chronic running injury himself, he not only started treating people with the same problem but recorded 157 episodes of the “Overcoming Proximal Hamstring Tendinopathy” podcast to talk about it at length and bring in new clients.

About a year ago, he paid a developer A$2,000 (about $1,300) to train a chatbot on some of those episodes and academic research on the injury,(1) then announced that listeners to his podcast could access “tailored advice” from a new AI assistant for A$8.99 a month.

Over the past year, around 50 people have signed up for the assistant, asking it about everything from the benefits of foam rolling to how to train for a half-marathon with the PHT injury. It’s making Sharpe an extra A$650 in monthly, passive income, he tells me, and after the set-up costs, he’s paid just A$20 in the last year on data usage to keep it going. The revenue is no fortune, but the side hustle could yet grow.

Eventually, he’ll train the chatbot on the advice he gives to clients on Zoom calls. What if it tells clients something he disagrees with? “I’m ok with that,” he says, adding that the bot is designed to replicate premium research more than Sharpe himself. “Whatever it comes up with should be good quality.”

There is of course, a thin line between rugged entrepreneurialism and excess. Deepak Chopra launched a “digital twin” last year that costs a few dollars to use, and after 95 books, a global brand and vast wealth already to his name, that feels less like innovation and more like an exercise in scaling himself ever further. Then there’s Snapchat influencer Caryn Marjorie, who became a millionaire at 24 after launching a bot that charged $1 a minute. Over time, maybe it’ll be the wealthiest among us who have the privilege of speaking to human experts while the rest consort with AI clones.

Still, there’s no denying the world is fast getting hooked on chatbots. More than 700 million people use ChatGPT once a week, and the total number is likely over a billion when factoring in those who use AI assistants from Alphabet Inc.’s Google and Meta Platforms Inc. As people become more accustomed to talking to machines, it won’t be a stretch for some to use them as a bridge to their expertise. That could widen all that know-how and ease workloads — if used wisely. Used badly, and they’ll turn more of our encounters into mostly cheap imitations.