Your people *Are* using ChatGPT: ensure it’s helping (not harming) your business
There is very strong scientific evidence that conversational AI, such as ChatGPT, substantially increases the skills, effectiveness and efficiency of your people.
But the impact of AI isn’t straightforward: it’s variable, uneven and uncertain. And it brings with it the potential for a surprising number of unpredictable and harmful consequences (which we explore below).
The critical thing to realise is that your people *are already using* ChatGPT.
How can I help
you today?
However, most businesses have no understanding of how they’re doing this, nor its impact (positive and negative). It’s shadow IT on steroids.
But do you want to wield its power consciously or let it impact your business in hidden, random and potentially undesirable ways?
For example, when its usage is hidden you can’t account for its impact. Are your people geniuses or just using AI? Are AI biases creating blindspots or gaps in your output? Is unsanctioned AI-assisted work causing resentment?
And using AI well requires conscious, strategic experimentation. As other businesses determine how to harmoniously weave it into their operations, you will get left in the dust if you take an unstructured, ad hoc approach.
You need to proactively understand and manage how your people use AI to ensure that it is helping (not harming) your business and to keep up with the competition.
How conversational AI impacts performance
While AI positively impacts skills across a wide range of fields, the rising tide of AI doesn’t raise all boats to the same degree.
This manifests in two main ways:
FIRSTLY
AI as a leveller
AI tends to level the playing field: low performers see much higher performance gains than already high performers.
For example, in a study involving a group of consultants from BCG, researchers found that the bottom half (performance-wise) increased the quality of their work by 43%, compared to just 17% in the top half.
Another study, by Noy and Zhang of MIT, found that half the performance gap between good and bad writers can be effectively eliminated by using ChatGPT.
SECONDLY
AI as a polariser
While in some ways, AI levels the playing field, in others it tends to warp it.
Using AI to improve your work is a skill and a talent. As in any other domain of human endeavour, some people are more naturally skilled at working with AI (or more naturally drawn to using it) than others.
In this sense, some people may only see small gains from using AI (because their personality or work style isn’t suited to it), while others will see massive jumps in performance. There’s also the concern that—as AI raises peoples’ effectiveness—it will simply make them lazy and dull, relying on AI to do their work for them.
So while some will be supercharged by AI, others may struggle to get their head around it or become over-reliant on it, making them less effective overall.
Key considerations for leveraging AI effectively
We don’t yet know how these varied possibilities will play out.
The potential upside is massive. What other technology can produce a 43% increase in performance with practically no overhead? But there are a surprising number of hidden and potential downsides.
Let’s look at some of the key nuances and considerations when it comes to getting the most out of AI.
Expectations will change
AI is going to change what you expect from your people (and even vice versa!).
If your people can work more quickly and effectively using ChatGPT, do you start expecting more from them? Or do you try to maintain the same output but by hiring fewer people?
Should you expect or even demand that your people use ChatGPT to be more effective?
The ethics of AI accessibility
It’s unlikely that everyone in your business will have equal access to the most cutting edge AI capabilities.
Business licences for AI tools like Microsoft Copilot are unlikely to be free. Are you going to buy one for every single person in your organisation? Or only those who need it the most? How do you judge that?
What are the ethical and social ramifications of giving it to some and not others? Especially if some are able to use it to supercharge their performance? Do you consider that in salary review conversations?
Social and political factors will remain
Although AI can massively augment skills in certain areas, knowledge-based roles are deep and multi-faceted, including many important social, cultural and relational skills.
These more human issues are foundational to a successful business but much less amenable to the rationalistic logic of AI. For example, a manager who lacks social skills or political adeptness isn’t going to improve these by 43% using ChatGPT.
Data security
In many cases, the data that you give AI to work with become part of the datasets they use to train the models.
With ChatGPT, this can now be disabled but if you haven’t trained your people to use it appropriately, they could be putting sensitive or customer data out into the AI ether, which may be a breach of data regulations.
The ‘jagged frontier’
AI is astonishing at certain tasks and lacklustre at others.
And where this so-called ‘jagged frontier’ sits is still being determined. This means that it’s not yet possible to say with confidence that AI will be able to help in domain X but not domain Y.
Businesses and their people will have to experiment within their unique context to see where it is most helpful and where it is least. Again, this kind of experimentation is a skill that some will be more adept at than others.
The question of AI is only going to intensify. The CEO of OpenAI has stated that there are developments in the pipeline that will make ChatGPT look like a “boring toy”.