Probably not yet, but it could with a few tweaks.
OpenAI released ChatGPT only a few months ago, and it's fair to say that it took the world by storm: it has over 100 million active users already. No wonder, when you can generate grammatically correct and human-like responses. Related technologies can also produce artwork and code by entering a description of what you want, and the technology produces it.
You can even interact with the AI after your initial question, so if you don't like the result you got or need clarification, you can ask additional questions or make adjustments to your image or code to more closely match your vision. All of this happens instantly without the help of a subject matter expert, artist, or coder.
But none of this comes without problems, which include the source of the data used to train the underlying AI model, the recency of that training data, the lack of permissions to use the source data, bias in the model, and, perhaps most importantly, the accuracy of the answers, which are sometimes ridiculously wrong.
None of this has stopped enterprise software companies from diving into generative AI. These companies see huge business potential and a lot of user enthusiasm and clearly don't want to be left behind.
Salesforce, Forethought, and Thoughtspot recently announced betas of their own flavors of generative AI. Salesforce is adding generative AI across the platform. Forethought targets chatbots and Thoughtspot wants to use AI for data query. Each company took the base technology and added some algorithmic drivers to tailor the technology to their unique platform requirements.
Microsoft also announced that its OpenAI service aimed at business users on Azure is generally available as a managed service.
Throughout this year, you can expect to see many more companies join in, but the limitations are real, which makes us wonder: Is technology, as early and raw as it is, no matter how cool it looks at first glance, really ready for the company?
A look at the limitations
Enterprise customers are grappling with whether to start using generative AI for business purposes but there are many unknowns.
The technology, as currently constituted, uses data to train the models without permission from various sources on the web, including text from websites, books, and articles. That's a big problem for everyone, but it's especially problematic for companies that create content for commercial purposes.
Marc Benioff, in an interview with journalist Kara Swisher at Upfront Summit earlier this month, noted that this is an obvious flaw, but it didn't stop Salesforce from releasing Einstein GPT last week.
“We can all see that ChatGPT is exciting, but we have all seen what the limits are. He is also the ultimate plagiarist. All the things you are learning you got from someone else. So your limits are the limits of the content you are trying to capture,” the CEO said at the time.
What's more, the answers are sometimes blatantly false, or at least partially incorrect. OpenAI even acknowledges this in its list of limitations of the technology, writing: “ChatGPT sometimes writes answers that sound plausible but are incorrect or nonsensical. Solving this problem is a challenge…”
Deon Nicholas, CEO and co-founder of Forethought, considers wrong answers to be one of technology's biggest problems. “ChatGPT will continue to blow your mind, right? If you ask him a question about a specific business, if he doesn't know the answer, he'll just come up with something that sounds plausible, but is completely wrong,” said Nicholas.
Also, ChatGPT is only trained on information up to 2021, which is problematic for companies trying to create the most up-to-date content.
There is also bias, which can be a real problem, and it takes a diverse team and careful attention to model and training data to help mitigate it. In a conversation on A Few Good Minutes with Brent Leary last week, Neha Bajwa, director of product marketing for customer experience at Microsoft, spoke about the importance of considering bias in AI.
“At Microsoft, we call it 'responsible AI,' ethical views on it and being able to do it responsibly, being able to make sure the data is unbiased, because [paying attention to] bias and inclusion is such an important thing. And data can amplify bias,” Bajwa said on the show.
How could it fit?
These limitations are not insurmountable. Software companies that have recently released generative AI tools have adapted the underlying OpenAI technology and made it their own to try to address some of these issues, but for now they remain.
Tim O'Reilly, who is the founder, president and CEO of O'Reilly Media, sees ChatGPT as the true third wave of the web, but says some adjustments will probably need to be made to meet the business requirements of the owners. of content.
OpenAI CEO Sam Altman even approached O'Reilly to train the corpus of knowledge in O'Reilly's catalog of books, but O'Reilly objected because it would require some kind of payment mechanism for authors who still does not exist.
"I said no until you have ... some form of payment, because this is a body of content and people expect to get paid for it," O'Reilly said. He suggested a system in which users would have to pay a fee to access this specialized content.
“Those [payments] would flow through the people who own the source content. Maybe we'll get a business model for these things where you have access to this more authoritative content,” he said.
One of the strengths is the breadth of AI's generative abilities, including text, art, and code. Nicholas said that being able to code workflows, as well as automatically create or adapt workflows on the fly, could be very powerful for companies that adopt this technology.
“One thing I will add, which is perhaps not obvious here, is that you can also use the generative models like GPT-3 to generate code, so you could also use them to generate workflows [on the fly], which is also pretty clean. So it's not just that we have an AI model that can talk and think like a human. But we have seen that GPT-3 can generate Python code,” he said. And that could lead to automated workflows.
Dries Buytaert, CTO of Acquia, who also founded the open source content management system Drupal, recently wrote a blog post about the possibilities of generative AI for content management and business in general.
In a recent conversation, Buytaert compared the development of this technology to cloud computing, which has fundamentally changed enterprise computing by providing quick and easy access to computing resources.
“The OpenAIs of the world are not just creating products, they are democratizing many of these tools. It allows a lot of people who don't have PhDs in machine learning and artificial intelligence to build very useful things very quickly. And that's pretty exciting, [limitations] and all that aside,” she said.
Buytaert suggests that, at the very least, companies should show their work and how they arrived at the answer. “They absolutely should give credit, and I think it would take away a lot of the discussion about organic search traffic impacts, for example, because imagine you ask a question, give an answer, credit your sources, and include links to the sources,” he said. .
That would be a start, and something that You.com, a search engine startup, already does with many of the responses in its chat-based search.
Bias is a more difficult problem to solve, but Microsoft's Bajwa says it will require a concerted effort by companies. “There always has to be human supervision. Technology can only help you so far. At the end of the day, [there needs to be] organizational structure and processes and governance that needs to be put in place because the technology is here to help and help the organization. The business has to set some parameters and recommendations and processes on how to use it,” he said.
That is precisely what any company will have to think about when looking at generative AI. As promising as it is, you can't just let it go and forget about it, and think you're going to get all these immediate results without any consequences. It is essential to keep humans involved because this technology is still too immature to leave it alone and hope for the best.
