Generative AI is unleashing big-time disruption to the ways of Business and work. New generative tools take assistive technology to a new level. ChatGPT has already made waves, raking in one million users in just five days. But enterprise uses for generative AI go much beyond ChatGPT.

Here is a list of generative AI tools beyond ChatGPT, already widely used.

• Murf, for text-to-voice applications.
• AssemblyAI, for audio-to-text.
• DALL.E-2, Open AI’s text-to-image tool.
• Stable Diffusion, another free text-to-image tool. 
• Synthesia, for text-to-video applications.
• StockAI, for AI-generated photos.
• LexPage, an AI-enhanced word processor.
• Interior AI, an interior designing tool.
• PatentPal, a legal process tool that automates documentation for filing patent applications.
• Jasper, a marketing content generator alternative to ChatGPT

These tools represent the upstarts of the nascent generative AI technology. As the technology develops and the industry churns, some of these tools will fade away, and new tools will unlock even more disruptions. 

AI-powered search

Many generative tools are in various stages of development. Most notable among the lot is Google’s Bard, which will become available to the public by the end of February 2023. Bard will take search to the next level by enabling AI-powered search. Users will get the benefits of auto-generated content along with search links. ChatGPT 4.0 will also be a big improvement to the existing Chat GPT 3.5. Microsoft will likely integrate ChatGPT with Bing and Edge browsers to compete with Google head-on. 

Products such as ChatGPT and Bard will augment search but not eradicate it. Only some searchers are content with a single response. As an indicator, only one percent of searchers use Google’s “I’m Feeling Lucky” button that serves up a single answer to a query. Also, most searchers seek answers from reputable sources and not from AI bots who can “hallucinate” up answers.

Content creation

Generative AI creates new, original content based on input parameters.

ChatGPT is an early foundation model. Presently, less than 2% of marketers generate content using ChatGPT or similar tools. But as the algorithm gets more mature and reliable, the use of generative AI for content creation will rise. Gartner estimates large enterprises will generate 30% of their outbound marketing messages synthetically, by 2030. 

Generative AI will find increasing use cases in: 

• Marketing. Generative AI will create customised campaigns, enabling more targeted marketing efforts. 
• Creating rich content. Generative AI will expand from written text to rich content such as videos. By 2030, AI will likely generate about 90% of major blockbuster films.
• Create original art, music, poetry and other literary works. Media and entertainment companies can create new content at a much faster rate compared to human effort. Such content will resonate more with the audience as the algorithm considers what worked in the past. Comparable human observations are superficial.

Generative AI’s ability to create large amounts of content improves enterprise efficiency. Businesses will leverage these tools to generate personalised content in double-quick time. 

Businesses can also leverage generative AI to unlock possibilities not viable before. For instance, they can use ChatGPT or similar tools to create personalised education materials. Human content creators will no longer have to stress over generating copious amounts of content. Their role will change to enforce supervisory control over machine-generated content.

Health care and drug discovery

Over the last three years, venture capitalists have invested over $1.7 billion in generative AI solutions. The bulk of the funding has been in AI-enabled drug discovery and AI software coding domains. 

Generative AI is adept at creating:

• Hypotheses and ideas for medical research. By 2025, generative AI technology will discover more than 30% of new drugs and materials.
• Generating synthetic data over empirical data. For instance, generative AI models can create healthcare data, based on the algorithms scouring through patent logs, without revealing the patient’s identity. 

The spin-off benefit of applying generative AI in drug discovery is substantial cost savings. A 2010 study estimates the average cost of developing a drug, from discovery to market, at $1.8 billion. Drug discovery costs represented about a third of the total costs, and the process took three to six years. Generative AI cuts the time to months. Pharma companies can reduce the costs and timeline of drug discovery.

Material science

Generative AI will find increasing use in designing products based on desired characteristics. It unlocks the creation of innovative products that are not viable using traditional design methods. Users may leverage generative AI to explore infinite design possibilities. These tools augment and accelerate design and invent novel designs or things humans may miss. 

Generative AI finds increasing use in the automotive, defence, medical, electronics, and energy sectors. The use cases include:

• Inverse design. AI-powered algorithms seek or discover materials needing specific physical properties for a product. For instance, a product may need more conductive materials than now. The incumbent methods are inefficient and time-consuming. For example, automakers can use generative design to innovate lighter designs and make cars more fuel efficient.
• Chip design. Generative AI’s reinforcement learning capabilities optimise component placement in semiconductor chip design. The product-development life cycle time reduces from weeks to hours. 

Generative AI enables the automation of several human tasks. Enterprise efficiency and workforce productivity improve, and costs reduce. But the technology displacing humans is still far away, if at all. Until then, human employees can focus their energies on more complex tasks, leading to job enrichment. They also get more time to reflect and make insightful decisions and engage with customers at a deeper level. 

Improving reliability of generative AI models

The love affair with ChatGPT’s uncanny conversational capabilities is unlikely to last. While cool, ChatGPT suffers from many limitations. OpenAI’s CEO Sam Altman himself admitted as much. He remarked that technologies such as ChatGPT are “impressive but not robust” when it comes to long-term, real-life use. 

Generative AI models generate responses but do not read sources or cite their work. Thus, the output has no depth and no guarantee of reliability. ChatGPT, for instance, always responds to a question, even if it is unsure if it is correct. The grammatically correct answer could fool an outsider but be complete nonsense.

The future of generative AI models depends on making the model reliable. The ways to do so depend on:

• Narrowing focus: ChatGPT is trying to be everything to everyone, from coding and writing poetry and essays to paralegal work. Narrowing down such tools to focus on a specific domain or task would increase the output’s quality, reliability, and depth. Generative AI could give reliable and factual answers to a limited domain. More so when the answers do not have a subjective element.
• Co-opting semantic search. Generative AI, in combination with semantic search, improves the credibility of the output. Generative models leverage semantic search “memory” to address the shortcomings of generative models. Semantic search enables priming the models with relevant primary sources and citing sources to the answer. 

Generative AI is still a nascent technology. It will take a considerable time for the capabilities of these systems to improve and make real-life mainstream use viable. Businesses would do well to stay up-to-date with the developments. They must constantly weigh the potential benefits and risks of adopting this technology.