How To Discuss AI Internally
Artificial intelligence may be a victim of its own hype. It’s certainly a victim of the various myths and rumors surrounding it. Dire prophecies from savants like Stephen Hawking, Bill Gates, and Elon Musk that AI is a threat to the human race don’t win the technology many friends. Even leading AI professionals, explaining why 87% of AI projects never make it to production, admit that: “What we are short on is trust.” And as Brad Power writes in the Harvard Business Review, you’re not going to be able to reap the full benefits of AI if your employees don’t trust it enough to use it.
Suppose, though, you’re a business professional who can see a solid case for introducing AI into your organization or enterprise. You have to win friends in the C-suite, to get approval for the RFP, and in the workforce, to avoid resistance and pushback against the plan. And you have all that prejudice and anxiety to work against. What to do?
One way is to manage expectations about the type of AI concerned, and what it actually is. No one is talking about advanced general intelligences, simply smart tools whose “intelligence” covers an extremely narrow range of tasks. Maybe AI tech vendors are the ones to blame for choosing the most dramatic label, but you need to demonstrate to your audience that narrow AI, or weak AI, is a genuine term among AI professionals, and that you are only looking at this type, which is very limited. This may make for a less dramatic, inspiring business case, but it also will make for a less scary one.
You should also hold genuine, open, respectful roundtable discussions and consultations with your workforce about AI and the proposed technology. If people like Bill Gates are concerned about AI, your workforce certainly has a right to have their concerns taken seriously. You may even want to hold a consultation period on the adoption of the new AI solution, or at least take soundings and make surveys to identify people’s concerns. This doesn’t mean you have to bow down to your staff, but at least make sure that your people know that their concerns have been listened to - and that you do know what those concerns are, so you can address them.
Make sure you do have information and examples ready on hand to answer common queries and allay common concerns, though. If people are worried about their jobs, hit them with the Gartner statistic that AI is estimated to have created more than 2.3 million jobs by 2020. If they’re worried about the dangers of AI, hit them with SAP’s use of AI to save lives in China through faster and more accurate diagnosis.
Another approach is to engage with internal units and divisions to make sure that the structures and relationships they’re comfortable with are supported, not eroded, by the new AI solution. Some change may be inevitable and desirable, but you gain better in-house support and more pro-active adoption if you involve your people in the development process. They’ll understand and appreciate how the AI can enable their work, and you may get useful insights into how it can best be applied.
For those who are simply afraid that their jobs will be automated out of existence, you can always point to manufacturing industries, which are still big employers regardless of the automation of their production lines. (You should probably also consult with HR to make sure that arrangements are in place to reassign anyone whose position may be made redundant by AI.) Emphasize that AI is about enabling better decision-making, not taking decision-making away. This will reassure your people about their degree of agency, and that they will still have control of their own working lives.
Learn to explain how the AI works in a simple, straightforward manner. This absolutely doesn’t mean explaining the details of algorithms and other such highly sophisticated components of the software. But if you can summarize and precis the overall operation of the system, even if it’s as a series of flowchart boxes, then you help your workforce understand the AI tool and become more comfortable with it.
Invite your workforce to experiment with the new technology, try their own solutions, and see what it can do. This not only builds trust and understanding, it also gives users back their sense of agency. And the more people feel ownership in the project, the more they will support it.
Reach out to a constituency that can benefit from the new technology, and build relations with them. This can be fully internal - one division or team - or it can be partly external, such as suppliers or even customers. Inevitably, there’s always a danger that a particular division or group within your organization will set itself dogmatically against the AI solution: when this happens, it’s good to have allies on your side.
By the same token, you need to handle issues with - and about - external suppliers and partners carefully. The new AI system may affect relations with them, but it’s important not to let them have the last word with your workforce. For your own people, you need to listen to how external relations may be affected, and what to do about it. Acting early on this could prevent a disaffected supplier spreading discontent about the new solution.
Build in benefits for your own people, and make sure you communicate these to them. AI often requires new skills from the workforce, as well as new technology, and your people will likely need those skills more and more in future. If you can give them better professional qualifications, upskilling, training in the new technology, or some form of career enhancement, you will have a workforce more committed to the solution.
If you can identify the key influencers within your organization who are most positive towards AI, give them the means to influence their colleagues by creating a knowledge archive about AI. This will give them (and you) a reservoir of good information and arguments to back up your solution, and will give you a team of authentic, convinced advocates, who will promote the solution on their own time out of sheer enthusiasm. The more info they have, the better they will argue on your behalf.
Obviously, a lot of these approaches are going to turn you into an AI evangelist within your organization - but that’s exactly what you should be if you want your solution adopted. If you have the commitment and the knowledge to support your solution, and communicate those effectively, you’ll get your people onside and embracing the potential of AI.