Artificial intelligence (AI) has reached dizzying heights, now touching almost every family of technology. AI is often used as a blanket term to describe a host of new advanced functionalities or capabilities, but not all AI is made equal. There are specific types of AI that fall under certain categories, such as deep learning (DL) and machine learning (ML), which have different capabilities and will have a varying impact on day-to-day life. One of the most talked about types right now is generative AI, as it probably comes closest to the sci-fi-laid expectation of what AI should be. As such, it’s making waves in how it can support all kinds of business tasks, processes and workflows. But what is generative AI, and how can it be used?
Whereas traditional AI focuses on decision-making and detecting patterns within pre-defined datasets, generative AI goes a step further to create new content from scratch in-line with specific prompts.
It’s important to distinguish generative AI from both machine learning (ML) and deep learning (DL). ML focuses on developing algorithms that learn from data to make predictions and action decisions e.g. facial recognition. DL is a subset of this and uses artificial neural networks to solve problems e.g. natural language processing. Although they’re closely related and often used in tandem, generative AI adds a creative layer through sophisticated modelling and more advanced algorithms.
A flood of new and intuitive platforms has seen generative AI rise to prominence.
Generative AI has multiple usages and capabilities, including the variant you’ve probably heard of the most – ChatGPT.
The 30th of November 2022 saw OpenAI launch ChatGPT, which gained 1 million users in its first 5 days, and 100 million users in the first couple of months. The website itself now boasts over 9 billion views. But for those that haven’t had a go yet, what can it actually do?
The best way to describe ChatGPT is as your AI doppelganger, thanks to its revolutionary ability to learn human interactions. People are increasingly using ChatGPT as their writing assistant to generate ideas, explain complex topics, and even write jokes. Business users are using it to help write complex Excel formulas and simplify their work processes.
But it’s not just language processing that generative AI offers. It can create song lyrics, music tracks, edit videos and form images based on different concepts. For example, Photoshop Beta’s Generative fill tool can take an image of a desert, and through adding the following text prompts: night sky, UFO, light beam, tree, road, puddle and car, it created this:
But beyond creating astonishing images such as that above, businesses can use generative AI to simplify projects such as enhancing customer service and technical support chatbots, manipulating it to process complex information and conduct a coherent, natural narrative with customers and prospects.
The major players all have plans to further maximise generative AI’s value within their product portfolios. HPE has announced future plans to accelerate generative AI and introduce it to their ProLiant Servers. Microsoft Copilot has placed generative AI front and centre in Microsoft 365 apps, and Microsoft Azure are promising generative AI experiences to modernize business processes. It’s everywhere.
But there are indisputable risks that come with these rewards. Generative artefacts can support increasingly complex scams, particularly image or video content created through ‘deepfakes’. Deepfakes can be created to mimic and manipulate almost anyone or anything, creating the potential for fraud and amped up cybersecurity risks through socially engineered cybercrime. We explore the risks of deepfakes further in this recent blog.
Because there are still so many unknowns, inputting personal or business information to developing systems such as ChatGPT could make you vulnerable to a data breach. Therefore, if your users are going to use generative AI platforms, mitigating controls must be put in place. Advice such as not inputting personal identifiable or sensitive information such as your company IP and outputs must be closely followed and monitored. It’s also important to see if your cybersecurity policy covers AI breaches. With the major vendors, you can operate slightly more trusting policies. One example is the Azure OpenAI Service, which adds security and compliance capabilities for generative AI, enabling businesses to unlock the true value of AI with extra security against cybercriminals.
It’s easy to get caught up in the hype of AI. But businesses will want to avoid AI washing, a newly coined term which refers to companies who falsely claim to be using AI. This can happen when a company doesn’t know what AI is, but they claim to use it because it sounds impressive, or they may genuinely think that they’re using AI but they’re not.
A recent study of European startups found that 40% of the businesses surveyed claim to be ‘AI startups’ but actually had no AI at all. Being known as an AI washout can have harmful impacts on your business, so if your business is planning to publicise its use of AI, just make sure you’re comfortable backing up those claims.
Generative AI is unlocking benefits to businesses worldwide. If you want to know more about what generative AI can do for your business, to explore some of the tools that may already be at your disposal, or indeed how to ensure the right cyber protection in an AI world, get in contact with a member of the team today.