AI trends in 2023: What to expect from the thriving field in the year to come

Artificial Intelligence (AI) is arguably the most ‘transformative’ technology to be developed to date. We are almost at the point in history where humans give birth to non-biological intelligence without evolutionary biases. The amalgamation of AI with cognitive science combines disruptive technologies like multi-modal networks, generative AI, machine learning at massive scale, etc. – that go beyond imitating the cognition and creativity of the human brain. Armed with contextual and spatial awareness, AI today can fix day-to-day problems and automate various tasks that were the sole domain of humans. 

Here’re some of the AI trends in this sphere that are going to take prominence in 2023:

AI creators

Text, speech, and vision AI will continue to come into the mainstream blurring the boundaries of human and machine creativities. So far, computers, software applications, etc. have been mere tools in the hands of a skilled creator for creating videos, articles, art, making decisions, etc. 2023 will be the year when the creator itself will be an AI agent. Generative AI output becomes indistinguishable from the human output. Varied applications in healthcare, mental health, coaching, teaching content generation, etc. will become predominantly AI-driven. 

AI co-workers in the enterprise workforce

AI is being adopted in organizations and the pace of adoption will skyrocket. Modern enterprises will be a combination of biological and artificial intelligence side by side. Enterprises will be onboarding AI co-workers that work along with humans increasing overall productivity and efficiency. In tune with these, the chatbots, conversational agents, etc. used in businesses will either transform or be replaced by AI capable of handling higher-order tasks requiring business decisions within the context.

CPA in the making – move from RPA to CPA – robotic to cognitive 

Yesteryears were the era of RPA which is now moving to AI-led automation or cognitive automation. Enterprises have realized the shortcomings of RPA in terms of brittleness, the total cost of ownership, and overall strategic benefits. With the maturity of AI automation platforms, a shift towards CPA is in the making, giving enterprises strategic benefits with AI and making them intelligent. 

Multi-modal neural networks

We are already seeing the benefits of multi-modal networks where the neural network is trained on different types of data (e.g. image and text, where the image is generated with textual input). Real-world data is multi-modal, (e.g. videos contain images and audio; a human can use sight, hearing, and smell to identify another person). These multi-modal networks open up a slew of business applications in domains including e-commerce, finance, manufacturing, healthcare, and more. 

Social value alignment for AI agents/workers

Social value alignment has been the forte of biological systems. A person might do something just because a loved one likes it without any other goal attached to it. How can the same happen with AI agents? That is, can an AI agent help with the goal of another AI agent? Each of the AI agents could be trained for different outcomes but can they coordinate for a common purpose? There is a lot of work going on in this space and 2023 could be the year where we start seeing value-aligned AI agents working towards a common goal for the enterprise. 

Explainable AI

Explainable AI becomes a reality for making ethical and fair choices while mitigating some of the enterprise risks associated with accelerated AI deployments. Modals will be thoroughly tested for bias and drift. Along with this, there will be an advent of data governance frameworks and tools to stay relevant and compliant with evolving legal and social structures. 

The article has been written by Sanjeev Menon, Co-Founder & Head of Product, E42