The advent of Chat GPT has taken the discussions of Generative AI to a broader sphere. Generative AI is now a subject of analysis even in average households. 

 To understand the current state of Generative AI and forecast the field's growth by the end of this decade, INDIAai organized a roundtable represented by experts from various fields. 

The panelists for the roundtable included Professor Chandra Sekhar Seelamantula, IISC Bangalore; Jyoti Joshi, Founder & CEO, Kroop AI; Emmanuel Goffi, Co-Director & Co-Founder Global AI Ethics Institute; Kazim R, Founding Director, The Dialogue; Aman Taneja, Principal Associate, Ikigai Law; Arvind Agarwal, Research Scientist, IBM Research India; Tapan Aslot, Assistant Vice President Marketing, IMS Group; and Abhinav Aggarwal, Founder, Fluid AI.  

Jibu Elias, Content & Research Lead- INDIAai, mediated the roundtable. 

Present status of generative AI? 

Generative AI has inspired many to gain first-hand experience with AI. However, an opaqueness in data collection, developers of the models, and biases being fed into the system still exists. This ambiguity will continue to remain even if the technology is evolved.  

The fear of being replaced by AI in the workspace has always been a significant concern. Generative AI is a catalyst in the escalation of this distress. However according to Jyoti Joshi, though there are concerns about AI replacing humans, in the present stage human intervention is a necessity. “Correct innovation is the key to the correct output", she added.  

Lack of knowledge in AI is yet another major issue. Different kinds of generative AI include images, texts, and language. "The public is lost in the complexity of the field", says Emmanuel Goffi. To simplify the concept, we are trying to streamline the systems, thereby failing to explain what AI is.  

In Emmanuel's opinion, explaining AI without hiding its complex nature is essential. Some people do not even have access to 4G in countries like India and Africa. Their knowledge of Chat GPT and Generative AI is limited. This clarity will lead to inclusivity and the maintenance of balance. 

Is AI creative? 

The roundtable became a platform for an open debate, where Prof. Chandra Sekhar pointed out that AI is still generative and not creative. In his opinion, creativity is still unique to humans as natural intelligence would be very hard to replicate. "AI is not completely well founded. That is the actual threat", he argues. 

Furthermore, Self-learning is still a dream for the models. However, with the AI models, the nature of jobs will change, and new jobs will emerge. But, for now, AI lack thought. The initial input still has to be from the human. 

Abhinav Aggarwal, who recently released a book written by AI, stated an alternative argument. Humans will also depend on reference materials before creating a particular model. "Why is that when a human does it, it is creative and generative when an AI does it?" he asks.  

Key takeaways- Ethics and Ownership 

Mentioned following are the key takeaways that the roundtable panelists remarked regarding the ethics in AI and Intellectual property rights: 

• The stakeholders involved with the work should be consulted to finalize the copyright claims.   
• In copyright law, there is a distinction between derivative and transformative work. While one is considered an infringement, the other is not, and these are very blurred and untested lines.  
• AI models largely depend on the data that goes into them. Therefore, data regulations have to become strict and stringent. Then, we will start having meaningful AI models. 
• User safety and content moderation will always remain an issue unless there are clear guidelines.  
• Roundtables to address the fear of academicians regarding generative AI is needed.  
• The ethics of AI should be addressed from the roots. Also, the concept of ethics can be subjective. Humanity is made biased. So, the question is, do we want to replicate humanity? Furthermore, who decides what is perfect and what is not?  
• The technology should be regulated based on its use case.