‘People feel AI hampers creativity and promotes plagiarism’

“The hype surrounding artificial intelligence, particularly ChatGPT, stems from its impact on various domains such as image recognition and behaviour analysis,” Muhammad Ali Jinnah University President Dr Zubair A Sheikh said during a recent dialogue session organised by the Ziauddin University.

“However, the concern lies in the negative effects observed, signalling the emergence of large language, image and speech models. This raises the question of control, as the internet’s openness and transparency have been compromised, suggesting that these models are being controlled.”

Dr Sheikh was sharing his views during the 19th Interactive Series on ZU Dialogues, titled ‘AI, ChatGPT, Google AI, OpenAI: Will artificial intelligence surpass human intelligence?’

The purpose of the dialogue session was to initiate a discussion on the significant challenges that confront human society as a whole. While utilising popular terms like AI, ChatGPT, Google AI and OpenAI, the focus was on addressing the pressing issues that have recently emerged regarding the future of humanity.

Acknowledging the widespread awareness of AI’s potential and the numerous benefits of information technology, the dialogue aimed to cover the concerns raised and delve into the complexities associated with these advancements.

“Recently, there has been widespread adoption of AI, leading to negative effects on the world, as evidenced by the emergence of ChatGPT. People feel that it hampers creativity and promotes plagiarism, among other concerns,” Dr Sheikh pointed out.

“However, this is just the beginning, as larger language, image, speech and even behavioural models are on the horizon. The question arises: are these models controlled by someone or do they operate in an open-ended manner?”

He highlighted the fact that the internet, although initially perceived as open, has proven to be controlled, with search results tailored to specific contexts and localised situations.

“As a result, the notion of the internet as an open, fair and transparent platform for information is no longer applicable. Thus, it is my belief that the large language and other models are being controlled by external forces.”

Unifonic Senior Country Director and P@SHA Vice Chairman Khurram Rahat pointed out on the occasion that whether we like it or not, AI has always existed.

“The crucial factor lies in how we utilise AI — whether it is used for constructive purposes or for negative areas. Just like the debate surrounding VCRs several decades ago, it ultimately comes down to how we choose to use the

technology.”

Rahat said that technology, including AI, is a tool that can streamline processes, assist individuals, and provide data and information to support human decision-making. It is a matter of perspective, he added, whether we see the glass as half full or half empty.

“When it comes to information about our preferences and thoughts, AI can offer recommendations, but the final decision to engage with that content remains in our hands.”

ZU Department of Computer Science & Software Engineering Chairperson Dr Munaf Rashid said that the debate surrounding AI surpassing human intelligence is an ongoing topic.

“Technologies like AI, YouTube, Facebook, Instagram and other social media applications are all integrated with AI. There is a high level of uncertainty about the future and whether AI has the potential to surpass certain aspects of human intelligence, but not all,” Dr Rashid remarked.

“It’s important to understand that human intelligence encompasses various aspects, beyond just image recognition or conversing with AI chatbots like ChatGPT. These generative models have limitations due to their database, which is generated by humans, and can result in issues like racism.”

He said that AI-based applications, trained on such datasets, may inadvertently produce biased or racist responses, which is a significant concern. Instances of AI making mistakes have also been observed, he added, such as the incident with the Tesla self-driving car that hit a pedestrian

despite detecting their

presence.

“While AI has surpassed human abilities in narrow domains like chess or image recognition, humour diagnostics, and even surgical accuracy, it is crucial to note that AI has its limitations and can’t encompass common sense, which is an essential component of human intelligence.”

The 19th ZU Dialogues, moderated by ZU Vice Chancellor Prof Dr Syed Irfan Hyder, concluded with the acknowledgment that the discussion held was highly valuable.

“As large language models continue to advance, a critical concern arises: the ability to differentiate between robot-generated and human-generated communication. This ambiguity raises complex issues, as it becomes increasingly challenging to discern whether we are interacting with a human or a robot.”