AJ Abdallat is CEO of Beyond Limits, a leader in artificial intelligence and cognitive computing.
getty
The rise of generative AI among the masses has been a complete paradigm shift. It unanimously passed the gimmick test—people have not just readily integrated it into their lifestyles, but they have gone as far as to expand its application to a range of queries that go beyond the scope of research and “ask me anything.” From using ChatGPT to plan a home renovation given a set of preferences, timeline and budget to remixing songs and producing original music using vocals from any artist as a blueprint, every day, we discover new potentials of this technology.
The layperson who uses ChatGPT to write a paper or get some direction before endeavoring to complete an unfamiliar task may form the loudest audience on this platform. But consumers don’t have to be the most voracious users of generative AI—businesses can (and should) also hop on board and harness its full power. It is imperative to look at generative AI outside the lens of consumer usage and avoid fixating on its limitations that don’t serve businesses. It is a malleable technology that works within your parameters and is often intelligent enough to extrapolate from those parameters when needed.
Generative AI offers boundless potential for businesses to optimize workflows, offer a superior customer experience and maneuver a range of business challenges.
Automation On Top Of Automation: Further Streamlining Of Processes
On top of the automation that AI naturally enables, generative AI has the capability to override software menus altogether. Users can work the software using direct commands without any pointing and clicking. The immediate advantage to this is even faster access to information and higher efficiency. But on a deeper level, it welcomingly blurs the line between the business and its technical inner workings.
Typically, a data scientist is required to encode business queries into computer language to generate business insights. With generative AI, the flow of command is shortened. It bridges that gap between any business users and technology—and that translates into a multitude of advantages. Enterprise technology in companies essentially becomes commoditized. Employees from several departments can readily use the technology to access relevant information that informs their decisions without intermediaries or having to navigate complex interfaces.
When companies are continuously finding ways to create accessible repositories of information—often among disparate and traditionally siloed departments—generative AI naturally facilitates that by further simplifying user interfaces. It whets the appetite to be more tech-savvy by removing the traditional characteristics of technology that business users have often resisted. The result is faster access to information for quicker decision making, which can significantly reduce time to market to keep businesses ahead of the competition.
Solving The Data Problem
Enterprise AI solutions can be even more effective and accurate using synthetic data. Synthetic data is artificially generated data that imitates real-world data but does not contain any personally identifiable information. Whether it serves as a complete replacement for real data or a supplement to it, synthetic data allows AI solutions to have a more robust data foundation to generate even more accurate decisions. Even when AI solutions can infer from existing data, that data can be inherently biased or simply not accessible due to privacy restrictions. Synthetic data opens the door for unchartered territories when there is simply nothing to build upon.
Cognitive AI can be a very useful tool in providing guardrails to ensure the synthetic data generated is realistic and consistent with the intended output. For example, if the intent is to generate an image of a scene, then the qualifying features you expect to have present can be encoded in a knowledge base. A computer vision machine learning algorithm can then analyze each generated image to extract what is actually present. A reasoner is finally used to compare what is expected versus actual and either accept or reject the result.
Fast-Tracking Code Generation
One of the most ubiquitous subsets of AI is large language models (LLMs). LLMs use deep learning algorithms to process and understand language. Neural networks create highly sophisticated LLMs, which allows them to analyze language so they can autonomously respond to queries.
LLMs are fluent in code, too. They can be trained in software documentation and generate code snippets or explanations of programming concepts and provide support for tasks such as code completion, debugging and refactoring. They can also be used to analyze and extract information from software code such as identifying patterns or detecting errors.
In that sense, businesses can use generative AI for fast, mass production of code, vastly truncating the time it typically takes for them to be created by large teams of programmers and testers. As with all forms of machine learning, large language models for writing code would need to be trained on huge amounts of data without being supervised. They would autonomously identify patterns and relationships in the code and build new ones from there.
Opportunities And Challenges For ChatGPT-Like Solutions To Marry Industrial AI
When it comes to scientific and engineering problems in the industrial domain, querying in natural language is not as straightforward as famously advertised by tools like ChatGPT. Typical enterprise search solutions can be tricky to navigate in this space because of domain-specific language, specialized applications and highly specific responses and actions. Whether it’s finding relevant information in an enterprise FAQ or knowledge discovery in scientific applications, it’s an area that still requires significant research—but there's also boundless potential.
The integration of generative AI into industrial AI is barely scratching the surface. There are applications in medicine, materials and formulation discovery, simulation, scenario generation, product research, development and design, among many others. The challenge lies in making tools like ChatGPT fluent in the language all these industries entail so querying is as seamless as possible. The amalgamation of both worlds of AI can lead to incredible buy-in for the adoption of industrial AI, which is often perceived as this larger-than-life technology that is too big, too complex and too elusive to trust. ChatGPT has made AI completely accessible; hopefully, industrial AI, which has lagged behind up to this point, will follow suit.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
