AI battle -Bard vs ChatGPT
Ravi Prakash
Senior Manager , Planning and Business Systems , Johnson and Johnson , APAC , MedTech
Have you seen the Hugh Jackman starring science fiction sports drama movie Real Steel?? If not then it is worth your weekend binging attention . What would surprise you most is the fact that film is based on the short story "Steel", written by?Richard Matheson, which was originally published in May 1956 . The story revolves around a boxer (Jackman) who himself does not fight . You would be wondering then why he should be called a boxer ! Almost a century back Richard visualized an AI battle which is unleashing right now in front of our generation but not literally in the same sense . Do we need to feel scared or enjoy this journey of evolution of a new specifies on our planet ? That is a million dollar question leading to concept of ethical AI . If we dream to explore vast unfathomed depth of universe around us then we must continue . Stephen Hawking had once opined that if Humanity has to survive then we should have another alternate home ! In this article , we bring two new kids on block to face each other and watch them engaging in a brave battle .
Round-1
Ah! We love cricket so why not start with IPL . Points table has become interesting now a days !
Question - Which team won IPL match yesterday?
BARD-
CG-
Verdict Round-1 - If you are a perfectionist then BARD wins this round as it gets the right information for us but CG 'honest ' answer is equally great ! Score -(B- 5 , C-4 )
Round-2
Before we start asking technical question which is their forte , we should engage them in easy encounters . What name they would recommend for my daughter?
Question - What is most common name of Indian girls ?
BARD-
CG-
Verdict Round -2 ----Bard gets the fact incorrect . Aaditya is common name of boy in India . But where it beats CG is getting data from Baby Center India. I like the response from CG though . If you are Indian then we know we have distinct common names in different part of country . Score - B -4 ( Because of data ) C- 4 ( Ability to understand the cultural diversity and then interpreting the answer accordingly . This is truly amazing) !
Round-3
Difficulty level should go up now . We have to keep this battle interesting .
Question - How would world look like after 20 years ?
BARD-
CG -
Verdict Round -3 ----Bard response is supported by data like Global temperature will go up by etc. CG has pretty much similar response and in fact more enriched if we ignore numbers . Score - B -5 ( Because of data ) C- 3.5
Round-4
Time to explore some technical questions . My daughter wants to know the weekday when her birthday will be as she grows. Let us see who can write a better python code to help her .
Question -Write a python code to know day of week from date ?
BARD-
CG-
Verdict Round -4 ----Both did a fair job but CG is better . Not only code is easily readable but interpretation it provides is like a human being . BARD has lot of ground to cover here Score - B -3.5 ( Because of lack of explantion ) C- 5
We can continue our game this way but this movie also must get over some where . Try some queries like -'Show me an image of elephant ' and you will notice that both BARD and CG would fail because these magics are part of specialized area of Deep learning which is called NLP ( Natural language processing ) while image classification etc. are studied under another domain of Deep Learning which is called Computer Vision!
Not getting into hype , you realize that both of these AI options are very powerful NLP stalwarts trying to exactly replicate human being but language is not statistics . Use both of these experimental capabilities as they bring different perspective ( Like that Python code. BARD code can be directly opened in COLAB) and that is the beauty of being human being .
With ongoing concerns about use of AI , i leave you with this beautiful quote -
'That is the end of movie'
Loved the way you challenged and evaluated both tools ????