Building a digital clone or assistant with generative AI — good idea or bad?

Disclosure: IBM, Microsoft and Cisco are clients of the author.

Generative artificial intelligence (AI) is moving at light speed, and we’re already beginning to see people use it to clone themselves digitally. (A Wall Street Journal writer created a digital clone that her family has been interacting with instead of her.) To reiterate how fast AI tools are progressing, the idea of creating a digital twin that could fool others was mere speculation a year ago, a sci-fi fantasy that was years, perhaps decades, in the future.

Not anymore.

A number of companies are working with generative AI now, so let’s look at the difference in, and danger of, these digital clone assistants and replacements.

IBM

IBM’s stated goal for AI is to supplement, not replace, people. The latest evidence the company is on this path is the recent announcement that it will pause filling nearly 7,800 positions as it assesses how AI might increase the productivity of its existing workers. To be clear, IBM is not laying off 7 ,800 people. It is in the process of determining whether the use of AI means IBM doesn’t need new employees. This is consistent with IBM’s stated policy and is not a replacement strategy.

Cisco

At a briefing this week, Cisco announced plans to use AI aggressively to manage and secure networks. With network admins and security personnel often overwhelmed, Cisco argues that using AI could reduce the workload on staffers by 40%. As with IBM, the plan is to supplement and enhance employees, not replace them. But it could help companies that use Cisco technology free up resources currently overwhelmed managing and securing networks.

Microsoft

Microsoft is integrating AI into every productivity tool it has. Since none of those tools act autonomously, this, too, is an enhancement, not a replacement approach. Microsoft’s move is foundational in that it will force users of Microsoft products to learn how to interact with AI tools and make those users more valuable as AI moves into more products. I see no sign Microsoft is looking to replace people with AI — at least not yet. It seems focused on using AI to help rather than replace users.

Smaller companies are also rushing to embrace AI.

Papercut

Papercut is building an AI-powered dubbing and video translation tool that can automatically localize any video with human-like voice-overs. The tool replaces voice actors who would otherwise be doing translations and voice-overs. This is more of a replacement tool because it means you don’t need an actual translator or a voice-over actor.

Literally Anything

Literally Anything’s tool allows a user to build a variety of digital apps, games, widgets and even services and could make graphic artists and game and app developers increasingly obsolete. This, too, is an employee replacement and elimination product, because if someone can more easily and cheaply use this tool for their creative needs, they would not have to hire someone else.

InBoxPro

InBoxPro appears to be more focused on enhancement than replacement as it provides capabilities that I expect Microsoft to eventually add to Outlook. InBoxPro works with Gmail to provide an assistant to manage email tasks, including contacts, calendar, sequences, and analytics. You might argue that such a tool could replace a secretary or an assistant, but those roles now involve many more duties, so this would be more likely to supplement than replace them.

Synthesia

Synthesia’s AI tool allows you to create a digital clone much as The Wall Street Journal writer did. This clone could be used to supplement or replace you, depending on how the company wanted it to be used. The question arises: who should pay for the tool? If you pay for it, you own the outcome and have a great deal to say about how your digital clone is used. If the company pays for it, then they own the clone and could eventually conclude that it can do your job just fine, making you redundant. This suggests a safer path is for you to pay for the service and tie the result to you instead of your company.

Who owns you?

The variation among different tech firms and their approach to AI suggests users might want to insist on owning any tool they’re training that could replace them. In contrast, tools such as those from IBM, Microsoft, or Cisco create little or no job risk and can continue to be purchased and owned by an employer. And tools such as those from Literally Anything and Papercut could make workers more valuable (at the expense of support staff who may no longer be needed).

As AI technology advances, we are going to have to think about what we want the outcome to be. I doubt most workers want to be replaced, but a tool that can take on mindless and repetitive tasks or make people more productive is worth using. The conundrum remains however: some people might gain value from these tools while others are in danger of losing their jobs.

In the end, the level of disruption we’re about to face is unprecedented, which means we all need to be clear-eyed and consider the personal implications of how our companies and coworkers implement AI and how we approach technology.