Generative AI washing: Avoid jumping the gun by laying the groundwork for successful adoption
Quite rightly, many businesses are excited about generative AI and the benefits it can bring. This year ChatGPT reported more than 100 million users, and the market value of generative AI is expected to grow year-on-year.
Currently, it makes sense that businesses want to get in on the act, but many are facing significant challenges navigating generative AI’s rapid emergence. There’s a risk of moving too quickly. It is difficult for businesses to confidently predict how beneficial generative AI will be. In some cases, it could even prove more of a hindrance than a help.
Jumping the gun?
There’s a risk that in their haste to adopt, businesses will end up 'generative AI washing' -- claiming they’re using the technology to offer better products or services, when in reality it is not having much of an impact, if it is having any at all. This could cause major reputational damage and lead to organizations falling behind competitors.
In a more general sense, AI washing has been a problem for years. Gartner has long-warned businesses about the risks -- and there’s now a similar problem with generative AI. To set themselves up for lasting success, there are three key areas organizations must cover before making any public announcements or launches related to the use of generative AI.
- Agree an overarching company strategy
Some businesses may be starting to use generative AI with a piecemeal approach. Others have instances of employees using generative AI under the radar. Both methods can cause problems, including different departments pulling in their own directions, or not learning from each other’s successes and failures. For example, an HR team may use generative AI to create job descriptions that don't fully meet the needs of the department that is being hired for.
Using generative AI too quickly could also pose security risks or compromise sensitive information. For example, if confidential information is entered into a prompt used to generate content, this data could end up becoming part of the generative AI tool’s training database. In turn, this could damage customer relationships or incur fines for breaching regulations like GDPR. The legal implications of generative AI have been under scrutiny and debate by the EU since at least April. Without an overall company policy to refer to, businesses could land in hot water.
- Lift the bonnet on your company’s tech stack
Generative AI cannot succeed without access to data. With the relevant data to hand, businesses can make strategic decisions about their choice of architecture, tools, and platforms needed to make a success from their use of generative AI. But first, they need to be confident they can access that data. If data is left in a siloed state, generative AI that’s used to automate certain tasks, such as inter-departmental queries, won’t be able to deliver the desired benefits.
There are other key considerations to look at across the tech stack, too. For instance, firms should consider how generative AI will be integrated alongside existing tools. Do they need to upgrade any legacy apps to enable it to work in practice? Generative AI is far from guaranteed to integrate seamlessly.
- Put the right skills in place
The third area for businesses to focus on is employee skills -- they may understand what generative AI is, but this doesn’t mean they’ll instantly be experts at using it. Inevitably, it will take time for staff to develop new skills – from setting up and integrating generative AI, to making full use of its outputs.
Depending on the scale of the activity, new hires with new skills may also be required, such as prompt engineers. These professionals engineer the prompts – or inputs – that generative AI uses to generate content. Recruiting people with these new skills may take more time and money than organizations have available, as the tech is still emerging, and experts are in high demand. For instance, Hays Recruitment recently observed that UK-based prompt engineers can expect starting salaries of £40,000, rising to a potential £300,000. It could therefore be more effective to look to partner with external consultants that possess these skills and bring a wider range of experience in harnessing them for different use cases.
Lining up your sights
Businesses are rightly enthusiastic to draw on the power of generative AI, but they shouldn’t go in all guns blazing. Especially with businesses that stand to benefit greatly from generative AI, the risk of reputational harm from rushed adoption is great.
It’s only once they’ve agreed on an overall strategy, found the right data architecture, and put the right skills in place to maximize their capabilities with generative AI, that a business can start to explore use cases and integrate it effectively.
Otherwise, they could be accused of ‘generative AI washing’ – with the time and resources they do invest ultimately generating little value.
Image credit: phonlamai/ depositphotos.com
Arun Ramchandran is President & Global Head -- Consulting & GenAI Practice, Hi-Tech & Professional Services, Hexaware.