Embracing generative AI early on can provide valuable business opportunities, especially when you have the tools to easily mitigate any associated risks.
In the future, as we will look back upon this time, we are likely to divide it into “before ChatGPT” and “after ChatGPT” - so big is the shift we are now experimenting as a society.
Just one year ago, the concept of generative AI (ChatGPT) was limited to a small group of forward-thinking business executives, and even then, it seemed like a realistic implementation was still a few years away. However, things have changed rapidly, and we now find ourselves in the midst of an explosive surge of interest in generative AI (ChatGPT) – meaning AI models that can create content based on given inputs. After all, the public version of ChatGPT alone attracted a staggering 100 million users within a mere two months.
Despite this surge in interest, large organizations and their CEOs have yet to fully embrace the excitement or explore the potential of generative AI (ChatGPT). In a recent survey of Fortune 500 CEOs, only 12% of leaders ranked generative AI (ChatGPT) as the technology with the highest potential for their businesses over the next decade. One would naturally assume that organizations would eagerly adopt any technology that simplifies daily operations and enhances employee productivity, especially one that requires no training or prior expertise. So, why is there scepticism surrounding generative AI (ChatGPT)?
For many top executives, there's no doubt that generative AI (ChatGPT) will eventually revolutionize their operations. However, the question they face is how quickly to adopt it at this moment. Some hesitation is justified because organizations are at different stages in their digital transformation journeys. Before jumping on the generative AI (ChatGPT) bandwagon, some still need to focus on getting the basics right, such as migrating their data to the cloud and establishing the appropriate environment and processes to facilitate secure and convenient access to it.
For others, though, their cautious approach is premised on different reasons. Here are the three top ones:
One major concern revolves around accuracy. People have witnessed the impressive capabilities of generative AI (ChatGPT) in delivering thoughtful responses, but these responses have sometimes been spot-on and, other times, not entirely accurate. Because of the way current generative models work, they can be prone to “hallucination”, which means they may provide answers that sound plausible but are unfortunately untrue.
In the business world, accuracy holds tremendous importance. Executives cannot afford to rely on a technological tool that is just as likely to fabricate information as it is to provide accurate insights. They cannot risk making decisions based on unreliable outputs from such a tool or having to be constantly double-checking the information it generates.
CEOs are equally concerned about the security implications of using generative AI (ChatGPT). Many are asking themselves, "Is it safe to rely on generative AI? Is the data shared within these models protected, and can we ensure compliance with regulations?"
Most experts agree that generative AI (ChatGPT) models are still in their early stages, and many of the tools available online are still undergoing training. This means that these models use the input they receive to further train themselves. While this can be beneficial, it also means that the information inputs are not fully protected and can be made public for training purposes. As a result, users may unknowingly leak sensitive data while using generative AI (ChatGPT) tools. One notable case of this occurred with Samsung in Korea, where employees inadvertently pasted lines of proprietary code into ChatGPT for bug-fixing purposes, leading the company to ban employees from using the tool.
Additionally, there is also concern that generative AI (ChatGPT) could be exploited by malicious actors to accelerate cyberattacks or manipulate outputs. Prompt injection, for example, is a technique where a third party provides new instructions to the model, tricking it into producing unintended outputs that may be malicious.
Currently, most generative AI (ChatGPT) tools, particularly those easily accessible to the public, do not guarantee data privacy. This poses a significant challenge for businesses that have contractual obligations to ensure privacy and confidentiality for their clients. When businesses input client, customer, or partner information into a chatbot, the AI engine may utilize that data in ways that cannot be reliably predicted. For instance, ChatGPT collects IP addresses, browser information, and browsing activities through cookies, which it may share with vendors or third parties without notice.
Another privacy concern arises from the fact that ChatGPT automatically opts users in. While the platform offers an opt-out feature in its terms and conditions for data collection, opting out may limit the type of answers users receive. Data privacy regulators, including those enforcing Europe's GDPR regulations, are still assessing these systems. GDPR's Article 17 guarantees the "right to erasure," allowing users to revoke consent and request the deletion of their data. However, chatbots currently lack the complete ability to remove individual pieces of data from their training, posing challenges for businesses subject to GDPR regulations. As a result, using AI bots may have long-term consequences for businesses under GDPR rules. In fact, Italy, as an EU member, has already taken action by banning ChatGPT due to concerns about potential GDPR violations.
In this case... is "wait and see" the best approach when it comes to generative AI (ChatGPT)?
Inaction stemming from fear is never a wise strategy, whether for businesses or evolving technology environments. In a rapidly moving global economy and a competitive market, even a decision to move slowly can have a substantial impact on a company's success and potentially leave them trailing behind competitors, making it challenging to catch up later on.
Aurelia Costache, EMEIA Intelligent Automation Leader at EY, gives straightforward advice for organizations wanting to start with Gen AI technology
Integrating ChatGPT into DRUID's Conversational AI platform through Microsoft Azure OpenAI services will amplify the power of conversational business applications and create even more value across a wide range of use cases. While we acknowledge the valid concerns mentioned earlier, there are practical ways to mitigate risks when the right approach is taken.
Moreover, it has the added benefit of being a plug-and-play solution – this revolutionizes the implementation process, offering businesses fast scalability, unwavering reliability, and seamless integration with Microsoft Azure.
The best thing, though, is that this integration ensures stringent data privacy measures, meeting the regulatory requirements of well-regulated industries while unlocking unmatched personalization possibilities for B2C interactions. This unique combination of robust data privacy and enhanced personalization helps businesses build trust with customers while delivering highly tailored experiences – and easily responds and puts to rest each of the concerns that most organizations currently have in relation to using generative AI (ChatGPT).
Alessandra Capogrosso, Data & AI Lead for Central Europe at Microsoft, explains the benefits organizations can access through a conversational AI platform infused with generative AI in a secure environment
Data privacy and security compliance are prioritized by utilizing robust measures already embedded in DRUID's Conversational AI platform to protect sensitive information, ensure confidentiality, handle data retention compliantly, and instill trust while meeting regulatory requirements.
A robust platform will help organizations design and test virtual assistants and Conversational Business Apps that deliver consistent and predictable responses in a business context, eliminating hallucinations or inaccuracies. This fosters reliability and usability, guaranteeing dependable outcomes and eliminating errors or misleading information.
DRUID prioritizes a transparent and manageable AI environment by addressing the challenge of the black-box phenomenon commonly associated with AI systems. It employs a set of transparent algorithms, processes, and auditing mechanisms, ensuring consistency, logical accuracy, transparency, and compliance. This approach facilitates troubleshooting and ensures smooth system operation, reducing the risk of unforeseen errors.
Integrated with enterprise systems, DRUID offers context persistence capabilities and leverages ChatGPT's generative AI (ChatGPT) power to initiate meaningful and engaging business conversations, effectively accomplishing enterprise tasks or fulfilling customer requests. This leads to the efficient execution of tasks and timely resolution of customer inquiries.
DRUID's advanced NLU capabilities for intent identification, working alongside ChatGPT, offer a powerful solution that aligns with specific enterprise requirements and goals. This will enable contextual understanding and effective operation in the business domain, delivering personalized experiences and improved outcomes.
DRUID's analytics dashboard empowers CEOs with the ability to make data-driven decisions and confidently assess the return on investment (ROI) of ChatGPT implementation. By providing accurate insights and measurable results, it enables CEOs to maximize the value and impact of ChatGPT, ensuring a successful adoption that aligns with business objectives and drives tangible outcomes.
The enterprise world’s reticence around generative AI (ChatGPT) is, after all, nothing particularly new. We've witnessed this pattern time and time again: an exciting innovation emerges, and businesses recognize its potential to disrupt the industry, but they hesitate to embrace it due to valid concerns that, in the grand scheme of things, may not be as significant as they initially appear. Going back in time, the same situation occurred around using third-party email servers or cloud services.
Of course, considering potential risks when making decisions in the business world is undoubtedly important. However, it is essential to also recognize that inaction often comes at a higher cost than the perceived risks themselves. History has proven that wasted time can result in losing ground to competitors who eagerly embrace change. Exploring the potential of generative AI (ChatGPT) from its early stages can be incredibly valuable, especially when you have the tools to easily mitigate any associated risks.
Businesses should get ready to enjoy the benefits of generative AI while realistically assessing the challenges that come along with its advancement.
Harnessing the power of conversational AI fused with generative AI can lead to unparalleled growth and lasting success for companies.
LLMs like ChatGPT have democratized knowledge work and creativity, shaping the future of work as we know it. But does AI augment or replace human...