Generative artificial intelligence (AI) is on everyone’s mind, with every sector of the marketplace – from banking and finance to healthcare and manufacturing – wondering how they can use this developing technology to their advantage.

The topic of AI carries a lot of questions and opinions. Generative tools like ChatGPT, GitHub, Copilot, and dozens of other music and art generators all come to mind. You’ve likely already dabbled in the technology, asking your friendly neighborhood AI virtual assistant what meal to prep for the week or how to draft an email to your boss. And you’ve almost certainly heard it discussed in the news. 

Emerging generative AI tools make it even easier to generate ‘new’ ideas, creative content, and solutions. But with that comes the risk of missteps with intellectual property, compliance laws, safety standards, data privacy and security, workforce impact, and fairness and ethics. Going forward, generative AI compliance regulations will play a huge role for industries interested in using this technology.

Generative AI poses a new possibility for reducing time-to-value in digital transformation initiatives by providing better accessibility of advanced technologies to the general population.

There’s a lot of fear surrounding the unknown aspects of generative AI, but how can we use it? And more importantly, how can we utilize it securely in our businesses?

What Is Generative AI?

To the untrained eye, generative AI appears eerily human. But it’s actually a sophisticated algorithm creating new, relevant content based on prompts and informed by the vast collated data on which it has been trained.

Generative AI is a powerful tool that works in various mediums including images, videos, music, software code, and even conversational interactions. But it’s not a replacement for people — far from it. Gen AI should be seen as a collaborator, not a terminator.

“Generative AI has a wide range of applications. For instance, it can be used for image generation, to create new music compositions, generate human-like text, enhance and modify existing content, or even simulate realistic environments for training purposes in fields like gaming or virtual reality (VR).

It's important to note that while generative AI can produce impressive results, it is still a developing field, and the generated content may not always be perfect or flawless. Nonetheless, it continues to advance and has shown great potential in various creative and problem-solving domains."

How does generative AI work? Generative AI utilizes various evolving techniques including natural language processing (NLP) and large language models (LLM), which are equipped to mimic human language. These applications are typically built on what we call foundational models (FM), which are AI systems based on deep learning. Think of them like training grounds where your AI studies and learns from a broad dataset. In its most basic form, AI uses predictive algorithms to generate 'new’ content such as images.

What Are the Use Cases for Generative AI?

Gartner sees generative AI potentially becoming “A general-purpose technology with an impact similar to that of the steam engine, electricity, and the internet”, stating that “the field of generative AI will progress rapidly in both scientific discovery and technology commercialization.”

It’s important to see generative AI technologies as assistants or collaborators. They add to human work through idea generation and research, while still leaving space for human empathy – something AI cannot substitute.

What’s a generative AI model? Typically, generative AI trains on deep learning models and then applies that to new content, mimicking what it learned from the training data. The generative models learn the underlying characteristics and distribution of the training data to generate new data samples that are similar to the original.

What do you use generative AI for?

Generative AI isn’t just about creating art, text, or music. It has the potential to transform day-to-day front- and back-office operations by streamlining workflows and aiding decision-making.

According to a McKinsey report, “Banking, high tech, and life sciences are among the industries that could see the biggest impact as a percentage of their revenues from generative AI.” They note the potential these industries have in automating the work activities that currently absorb 60-70% of their employees’ time.

Generative AI can be applied to a range of fields. Here are some high-level generative AI applications:

• Producing text drafts and outlines for written content
• Discovering answers more quickly through text analysis
• Developing AI-generated images for idea and design exploration
• Improving customer service assistance through chatbots with intent/query understanding and language processing capabilities
• Summarizing data-heavy information by extracting key information
• Simplifying complex documents
• Generating, testing, and implementing software code
• Changing the tone of a piece based on formality and context
• Sorting and organizing information based on specified parameters
• Increasing productivity and revenue by streamlining operational workflows
• Implementing automated catalog management for product lines and stock

Generative AI use cases in healthcare

The healthcare industry may be one of the largest sectors to benefit from AI and intelligent automation (IA). Generative AI has the life-saving potential to help clinicians reach an early diagnosis for their patients. For example, it can assist radiologists in detecting cancer spots by analyzing medical images for a more accurate, faster diagnosis.

Technologies such as natural language processing (NLP) can read large, unstructured healthcare datasets like you’d find in electronic health records (EHRs). This gives generative AI the power to convert X-rays, CT scans, and MRIs into more realistic images, helping identify potential anomalies easier.

As this technology develops its decision-making capabilities, it can use healthcare datasets to input a patient’s medical history, lifestyle, and genetics, helping create a personalized treatment plan. This improves healthcare patient services and gives more time to clinical staff to focus on patient care.

Generative AI use cases in life sciences

Pharmaceutical companies can utilize generative AI to analyze large data sets on side effects, clinical study results, and efficacy. As this develops, it will combine these larger data sets with smaller insights collected in real-time from patients and individuals within a target population, allowing it to make connections within the data and prescribe possible solutions based on previous clinical trials.

Gartner estimates that “By 2025, more than 30% of new drugs and materials will be systematically discovered using generative AI techniques, up from zero today.” This will have huge impacts on the pharmaceutical industry by reducing drug discovery time and possibly reducing time-to-market by creating synthetic data based on clinical trial information already collected.

According to Deloitte’s study of intelligent drug discovery, AI will “Open up a new future for the health industry as a higher level of knowledge on disease mechanisms both increases the number of treatments available and, in many cases, cures diseases that have not previously had effective treatments.”

Deloitte suggests leaders with digital knowledge will need to integrate new strategies into their research units as generative AI capabilities expand. To date, the U.S. leads the AI drug discovery sector, with 60% of startups utilizing the technology, while the usage level in the U.K. and European Union has increased by 10.8% and 10.3% respectively.

Generative AI use cases in banking and finance

With a good governance model in place, generative AI can help with risk management and fraud detection.

Finance services and banking are heavily regulated industries, where compliance is of the utmost consideration. Generative AI has some evolving to do before it fits into every aspect of banking, such as regulating the private document protection, but it has the potential to analyze data for credit scoring, portfolio management, trading, price optimization, market predictions and so on. Here are a few ways it does this:

• Fraud detection
• Anti-money laundering (AML)
• Financial services product recommendations
• Risk assessment and risk management
• Customer service via chatbots and automated scheduling
• Financial market predictions

Generative AI use cases in IT

AI-assisted software development is making huge headway in programming. With technologies like Copilot, generative AI can assist in code generation using a vast database of programming techniques.

With generative AI’s ability to synthesize data, it can execute tests within the system to check for errors or inefficiency, helping developers optimize code. Once the code passes inspection, AI can automatically deploy it into the workflow and continuously monitor it as it works.

Generative AI solutions can be utilized in software development by:

• Suggesting the next line of code
• Answering code-related questions
• Generating new lines of code based on text prompts
• Testing software functionality
• Identifying and fixing bugs in code
• Helping implement new models into software based on existing models

In truth, generative AI will likely touch every sector in the future, from automotive and aerospace testing to marketing messages and personalized services. It’s a bullet train of innovation with a lot of potential. But as with any fast-moving machine, there are risks.

What is prompt engineering? Developers who have achieved a good understanding of the generative AI model and its desired output can utilize prompt engineering techniques to experiment and fine-tune different prompts, optimizing the model’s performance to align with the intended task or objective.

What Are the Risks of Generative AI?

Maybe you’ve heard some of the risks of generative AI before: deepfakes, copyright infringement, leaked private documents, or publishing incorrect data. AI can confidently tell you the wrong information without skipping a beat, and deepfakes appear as convincing videos of politicians or celebrities saying words they’ve never spoken.

Technologies like ChatGPT and other generative tools are trained on public data with no copyright law or General Data Protection Regulation (GDPR) compliance. In other words, if you’re using generative AI in your enterprise, we recommend caution in how you approach it and what you claim as ‘yours’.

However, with the right models, guidelines, and oversight, generative AI has the potential to completely transform work in a range of industries, including healthcare and life sciences, manufacturing, banking and finance, energy, IT, and so on.

Bringing Governance into Generative AI

Generative AI is an unstoppable force, so let’s consider how we can use its powers for good while always considering the Environmental, Social, and Governance (ESG) impact.

Before organizations start rapidly implementing generative AI into their operations, they should consider its impact on their reputation, revenue, and potential legal liability. To do this, organizations will need an established governance risk management strategy, with rules and principles for how generative AI can support their automation strategy while maintaining compliance.

What governance does for generative AI:

• Protects laws, privacy, and intellectual property rights
• Aligns AI with your organization’s ethical principles
• Checks the accuracy of AI output
• Maintains your organization’s quality standards and consistency
• Ensures privacy protection and data accuracy
• Considers sustainability and environmental controls

Regardless of how we allow generative AI to impact us, it’s a technology that’s running full steam ahead. It can make a positive impact, so long as organizations use it with governance and data protection in mind. And with rapidly changing laws regarding generative AI, your organization must monitor the regulatory environments in their countries and spheres of influence.

What’s a large language model (LLM)? Large language models (LLM) are a type of AI model specifically designed to understand and generate human language, mimicking normal human responses. They’re trained on vast amounts of textual data and can perform various language-related tasks including text generation, summarization, query understanding, translation, and so on. Typically, LLMs are based on deep learning architectures. There are various applications of LLM such as chatbots, virtual assistants (like Amazon’s Alexa or Apple’s Siri), content generation, and language translation.

Implementing generative AI into your IA

There’s a lot to consider when bringing generative AI into your automation program. Here are some top tips you should include in your strategy:

• Establish guidelines: Define the ethical boundaries of your AI usage and where and how you’ll deploy it, also considering the impact on the individual, society, and privacy rights.
• Align with goals: Align your technology with your organization’s ongoing strategies – assessing, prioritizing, and managing how generative AI fits into your technical capabilities and business requirements.
• Identify use cases: Determine the specific tasks or processes in your IA workflow that could benefit from generative AI and establish what effects those integrations may have.
• Use quality training data: Carefully curate your training dataset, ensuring whatever data your generative AI trains with will fully represent your organization’s desired outputs.
• Validate content: Thoroughly validate your output to avoid incorrect data or biases.
• Enable the workforce: Provide comprehensive change management strategies with the people in your organization and establish best practices and goals.
• Maintain human-in-the-loop (HITL): Bring people into the automated processes, such as a Center of Excellence (CoE), to ensure best practices are maintained.
• Adapt: Regularly provide feedback and data to your generative AI model so it can adapt to changing requirements, continuously improve output quality, and expand for better scalability.

What Are the Benefits of Generative AI?

Generative AI’s capabilities and accessibility are unprecedented, giving the automation space in any sector the potential to use and benefit from it. Advancements in natural language processing (NLP) especially will come into more focus for industries such as healthcare, finance, and customer experience (CX).

Here are the key benefits we’ll see in future uses of generative AI:

• Content generation
• Personalized experiences
• Problem-solving
• Decision-making
• Simulations and testing
• Data augmentation
• Creative content experimentation

How will generative AI provide value to businesses?

Generative AI can provide businesses with the opportunity to:

• Increase productivity and revenue
• Reduce operating costs
• Improve risk management
• Ease employee workloads.
• Work quickly and efficiently
• Improve scalability, consistency, and quality

With this comes the caveat that a generative AI system only knows what it’s taught – and with restrictions well on their way with copyright protections and ownership laws, it’s a technology that organizations must use with caution, and with the backing of good governance and best practices.

Generative AI is a content creator, first and foremost, and that’s going to change the nature of people’s jobs. In fact, Gartner predicts that “By 2026, over 100 million humans will engage robo-colleagues to contribute to their work.” It’s up to organizations to anticipate this shifting landscape and look at how they’re going to react to the future of work.

With the rapid advancement of AI and the potential it holds for transforming industries, it’s important to partner with reliable and innovative solution providers like SS&C Blue Prism.

What’s the Future of Generative AI?

We interviewed ChatGPT to find out what generative AI was all about (of course, verifying the truth of those statements with further research). And with every response, ChatGPT came back with a similar, final warning in several iterations. Here’s one of them: