.png?width=180&height=65&name=Ascertus%20Logo_360x130%20(1).png "Ascertus Logo_360x130 (1)")
Categories:
AI
There is such excitement about generative AI – it’s the ‘real deal’ they say and so, many organisations appear to be in a rat race to deploy it. Of course, its benefits are irrefutable and in the near to medium term, we will start to see it being meaningfully deployed, just like some other historically game-changing technologies now are – email, mobile, and cloud are examples.
AI is here to stay, but approach with caution
Generative AI is fresh out of the labs and needs to be honed and fine-tuned so that glitches, hallucinations, discriminations and biases are ironed out, so that it delivers real value to the user base. If some of those related terms are new to you, all the more reason to follow this advice.
The data that generative AI is trained on is the most important, and so any organisation that wants to adopt the technology will do well to cleanse and properly structure its data. If the data is internal and of good quality, the quality of output will be high and relevant, making it accurate and reliable for users to safely use.
Software vendors such as iManage are already pursuing this path with 3 new iManage AI features due to imminently land. An organisation’s document management store is the most logical place to start for the adoption of generative AI. With all the firm’s content residing in a single location, embedding AI in the document store means that the model is automatically trained on the thousands of documents that the organisation creates. This also means that users can easily search for specific information or knowledge and seamlessly apply it to their day-to-day workflow. It reduces the time it takes users to find the specific information they are looking for, and focuses on the value-added activity that is critical to the successful delivery of the tasks at hand.
Generative AI - you can't always see it
Most crucially, in such a scenario, ‘generative AI’ technology becomes invisible to users. When this happens, it’s safe to say that the technology is truly well-deployed and mainstream. To draw a parallel, we all use Google and Amazon, do we really care if these platforms are using AI or something else? We just want intuitive search so that we find what we are looking for, quickly and safely.
Of course, there are use cases that the wider industry hasn’t even thought of yet. Customer support may be one such use case. Tools such as ChatGPT can potentially help with delivery of better customer and brand experience. Many websites today already have chat bots embedded, but their effectiveness varies. A generative AI tool may help significantly expand functionality.
There’s no doubt that the evolution of ChatGPT/generative AI has been exceptionally rapid since it first burst on the scene a few months ago. Application vendors are working against the clock to safely incorporate the technology into their software and it’s in the interest of all of us as users to afford them the time to do so, in order to derive tangible value and ROI. Some like iManage are making tangible headway – the recently announced new iManage AI engine for document classification and enrichment is an example.
Until then, caution is advised. Language is a beautiful thing – one can say something in so many different ways. I’ve read a few articles generated by ChatGPT, and it strikes me that you can also say nothing in so many different ways. The sketch where Eric Morecambe plays the piano, and the world famous pianist and composer, André Previn, tells him off for playing poorly, comes to mind. Eric then insists, “I played all the right notes, but not necessarily in the right order”!
See also
FAQ
What is the biggest challenge facing AI adoption?
The biggest challenge in AI adoption is trust and ethics. Key issues include ensuring data privacy and security, addressing bias and fairness, and improving transparency and explainability of AI systems. Navigating evolving regulations, managing ethical concerns like job displacement and surveillance, and overcoming technical integration challenges are also crucial. A multidisciplinary approach involving technologists, ethicists, policymakers, and public engagement is essential to create trustworthy, transparent, and beneficial AI systems.