Why manage your AI?
- 4-Minute Strategic Read
- Kenya Data Protection Act (ODPC)
AI has been used in business for decades. For years, it was the preserve of data scientists and engineers on large, leading edge IT projects that delivered things like self-driving cars, online shopping recommendations and semi-automated credit scoring.
That all changed in November 2022 when OpenAI released ChatGPT, a conversational AI that everyone could use, reaching 100 million active users within 2 months. Since then, the technology behind ChatGPT has been offered by a variety of suppliers such as Anthropic (Claude), Google (Gemini), Microsoft (Copilot) and others. More importantly, it has become embedded in just about every software application or web service that is used in business, whether you like it or not. The AI-written answer at the top of your web search, the “help me write” button in your email, the meeting summary that appears after a Teams or Zoom call, the Meta AI assistant inside WhatsApp – these are all generative AI at work.
Why this matters
Modern (generative) AI is a powerful technology. But it is just that, a technology. At its core, it is a prediction engine – it looks at what you entered (the prompt), matches it to what it has seen before (its training) and gives you what it decides is the most likely output (the response). Probability plays a role in how these AIs work and so if you ask the same question twice, you may not get the same answer. For example “What is the capital of France” might give you “Paris” 999 times out of 1000, but one time it might respond “F”, as in the capital letter in the word.
This makes it fundamentally different to conventional computer technology. With conventional software, you test it and it provides an answer. Repeat and you will get the same answer. AI does not work like that. Worse, it can give answers that sound plausible and confident but are simply wrong – often called “hallucination” – and users need to be aware of this. AI can be wrong and if you are making business decisions based on what it says, you need to have checks in place.
It’s your data, until it isn’t
Data protection is something we are all aware of. Many companies have invested heavily in technology to store and manage sensitive data, whether it’s customer information, regulated information such as health data or just the business’s confidential IP. But when you put data into an AI, unless you have taken very special measures, that data will go somewhere you don’t control, probably overseas, and what happens to that data depends on a licence agreement that at best you may have clicked through.
More importantly, even if you have all of the applications in your business that contain AI (and there are probably dozens) well controlled, there is nothing to stop an employee pasting the latest sales numbers into a free public AI via their computer or their phone, in order to save time on creating the monthly sales report. And most free AI tools store what you give them and, unless you opt out, may use it to train future models, which means it is now out there in the wild
Cyber on steroids
AI is a powerful tool. For adversaries as well as businesses. Attacks that used to take days now complete in minutes. If there is a vulnerability in your system, AI dramatically increases the chance that someone will find and exploit it. In addition, the AI in your business is like your personal bodyguard. It’s your most trusted team with access to the castle keys – but if someone is able to subvert that AI (a technique often known as “prompt injection”), the AI can be used against you, for example to quietly copy your customer database out of the business. Knowing what AI you have and what it has access to is key.
This matters even more with AI agents – AI that doesn’t just answer questions but takes actions on your behalf, such as sending emails, updating records or making bookings. Agents are increasingly built into everyday tools, and each one is only as safe as the access you give it. This is not an exact science – an agent that can read your inbox and send email can be tricked by a single malicious message into forwarding things it shouldn’t. Agents, while powerful, add a new layer of complexity to potential AI harms.
Don’t be afraid, be prepared.
Like it or not, AI is with us. It is already all over your business and your employees have a dizzying array of approved and unapproved services they can access. The three steps you need to take are:
- Know what you have
- Decide how you are going to manage it and put that into practice
- Train your people to know how to use AI safely
- Key Takeaways
- AI is already built into the tools your business uses every day, whether you chose it or not
- AI can be confidently wrong, and the data you give it may end up somewhere you don’t control
- Start by knowing what you have, deciding how to manage it and training your people
To explore further
Take a look at our Resources page with many other helpful articles on AI and governance
Try our tools and solutions
Discover Akili’s tools and solutions designed to help you understand,
assess and govern AI with confidence.