Artificial intelligence has endowed businesses with limitless capabilities: process automation, data-driven decision-making, and customer and employee empowerment. In its next act, AI will evolve from augmentative technology to a more direct generator of products and data via Generative AI.

The new generation of AI is Generative

Generative AI refers to artificial intelligence algorithms that foster the creation of original materials from existing content such as text, audio recordings, or images. In another way, it enables computers to extract the underlying pattern associated with the data input, then subsequently generate similar materials called synthetic data. While generative AI presently accounts for only 1% of all data produced, Gartner predicts that by 2025, it will account for 10% of all data produced. Synthetic data, on the other hand, will account for 60% of the amount of data used for AI and analytics solutions in the next two years [1]. An AI-driven future is filled with endless possibilities.

If deployed effectively, synthetic realness can push AI to new heights. It can deliver next-level improvements to AI models in terms of fairness and innovation by addressing data bias and privacy issues. AI-driven robots are structured with the ability to comprehend more abstract notions in the real world, resulting in a better generation of authentic, life-like materials. Consequently, such synthetic content will allow customers and employees to experience more seamless, innovative AI experiences.

Many businesses are thrilled to jump on this "AI 2.0" bandwagon for its noticeable impacts. According to a survey by Mckinsey, modern deep learning AI techniques could deliver an increase in additional value from 30% to 128% beyond traditional analytics techniques, depending on the industry [2]. For example, in the manufacturing industry, 30% of manufacturers will apply generative AI to increase product development efficiency by 2027 [3].

Use cases of Generative AI in various sectors

With such significant benefits, various manifestations of generative AI have already taken off or are forecasted to:

  • Healthcare: By 2025, 50% of drug development initiatives will use generative AI [4]. Generative AI enables early identification of potential diseases to produce effective treatments while the condition is still in its initial stage. For instance, AI computes different angles of an x-ray image to visualize the possible expansion of a tumor. Synthetic datasets that simulate the original information are published for external access instead, consequently protecting the patient's data privacy.
  • Marketing: An early implementation of generative AI technology lets companies plan marketing content with a higher engagement rate - centered on improving personalization and campaign performance. Concerns over AI's potential to sap the human touch from the brand's voice could be quelled by contents delivered by hi-tech bots. This content showed higher click-through rates than a traditionally generated copy. By 2025, 30% of outbound marketing messages from large organizations are expected to be synthetically generated [4].
  • Media & entertainment: Machine learning algorithms and generative AI can create authentic dubbing speech in various languages using AI-driven deepfake technology. Apart from that, generative AI is also useful for increasing the resolution of images and videos in motion pictures, upscaling them to 4k and beyond; generating more frames per second (e.g., 60 fps instead of 23), as well as adding color to black and white movies. Old films can now be revived with the same level of clarity and resolution as modern ones.
  • The metaverse: Artificial intelligence will provide fundamental support to the metaverse, such as simplifying people's access to digital environments and supporting content generation. Generative AI will be able to recreate any existing place in the world, generating stunning 3D scenes from still photographs.

But what are the risks?

  • Security issue: Risks of Identity theft, fraud cases or counterfeiting cases lurk as Generative AI allows the generation of seemingly-realistic photos and images. Disinformation campaigns might take advantage of the fact that it is challenging to distinguish deepfakes from the real. In 2020, a UK-based energy firm was hoaxed into sending roughly 200,000 British pounds to a Hungarian bank account after a malicious individual used deepfake audio technology to impersonate the firm's CEO's voice to authorize payments [5].
  • Data privacy concern: Since it involves collecting the privacy information of individuals, there is a considerable risk to individuals' rights and freedoms. This is quite different from the threat posed by data breaches and with very little "fallout" for the firms involved. Some of AI's privacy concerns include: Data persistence (data that lasts longer than the people who developed it, thanks to low data storage costs), Data repurposing (use of data for purposes other than those for which it was originally intended), and Data spillovers (obtaining data of people who are not the intended audience).

Businesses benefit from generative AI models because they can learn from themselves and generate fresh data, which is cost-effective and efficient. New use cases shall emerge over the next few years as more firms experiment with this innovative technology.

Author FPT Software