Since its launch in November 2022, the AI chatbot ChatGPT has gained remarkable interest, with over 100 million active users monthly. The tool has also sparked interest among business leaders for its ability to augment communications with clients as well as to assist with activities such as software development.
Developed by OpenAI, ChatGPT is an artificial intelligence program built on a family of large language models (LLMs). ChatGPT has the sophistication to generate a vast array of text, express creativity, and manage and respond to queries in a conversational style. As many users have already discovered, ChatGPT can handle complex tasks, from writing code and music to assessing its own flaws. In December, the bot became a mainstream sensation when it successfully passed the Turing test, making it only the second bot in history to do so.
ChatGPT has the potential to lessen human workloads while simultaneously enhancing customer experience, which could result in business growth. While early users have reported that ChatGPT and other AI-powered chatbots produce favorable results, the rewards must be considered alongside the risks.
Questions About Security and Ethics Continue To Evolve
Security analysts, business leaders, and even U.S. lawmakers have questioned what chatbots like ChatGPT and LLMs might mean for data privacy and cybersecurity.
Some of those questions surround ethical AI: Sophisticated chatbots could be used to generate and spread misinformation and hates speech, develop and disseminate phishing attacks, aid students in cheating, and even create national security issues.
In addition, chatbots today can struggle to process complex queries and respond appropriately in sensitive or nuanced situations. That can result in strange, potentially offensive, and factually incorrect chatbot responses.
Other questions surround the information that chatbots can access. Chatbots may gather and store customer data that could pose privacy and security risks if not properly handled.
When it comes to data protection, the best defense is for organization to manage the information it is exposing to public – and chatbot – consumption. “The No. 1 thing that I encourage everyone … is [to] make sure you have an idea about what information you have … but also [to] start thinking what it would mean to expose this information [to the internet],” said Nathan Howe, vice president of emerging technology at Zscaler. “It is nearly impossible to remove all traces of information from the internet once it’s out there.”
Only after an organization understands what information it’s making public can it put controls on the sensitive information it wants to protect.
While ChatGPT might be a fairly new advancement, the challenges surrounding security and data protection are well established, Howe said. Businesses can take the same steps to secure their data that they do for protecting against ransomware, data breaches, or phishing attacks: Educate staff on proper security protocols, keep software up to date, and use secure passwords and multifactor identification. Most critically, perhaps, is for businesses to draw a line between the information they share online and keep private and decide how they will ensure they use AI ethically.
The Future of AI in Business
Given ChatGPTs overwhelming success, it has become less a matter of whether the technology will influence businesses than how businesses should respond.
All industries are exploring how they could use generative AI. With companies like Google and AWS investing in rival chatbots, organizations should expect to see generative AI tools increasingly incorporated into their business processes. “For years, tech companies have used [machine learning] and AI in their platforms, but only now are people really beginning to understand what it means,” Howe said.
When weighing the benefits of adopting AI technology like ChatGPT, it will depend on the specific needs of an organization. However, despite AI being capable of augmenting (and in some cases replacing) human workloads, humans have a large role in ensuring AI tools like ChatGPT are used responsibility.
“It's very powerful if we … play with it the right way, but it can also be misused if misunderstood,” Howe said.
