ChatGPT - Perceived sides of the same coin

ChatGPT (My Sakha) - a technology that has been hailed as revolutionary, game-changing and even perceived as potential existential threat. With opinions on both sides of the spectrum, it's no surprise that CISOs, Chief Enterprise Architect, CIOs and other tech experts are divided on its use!

As someone who is a self-proclaimed power user ?? of AI-powered tools, including ChatGPT (& in tandem with other AI powered tools), I have had the opportunity to explore its benefits firsthand. Whether it's creating music and videos (predominantly 3d Animations ) for my customers, writing emails or generating code for my freelancing projects (eCommerce, ERP & others), AI Tools in general & ChatGPT in particular has become an essential tool in my arsenal.

Speaking about ChatGPT, I must admit that I adore it (and I believe I am not alone), whether it is used on the OpenAI chat website or other applications that use OpenAI APIs (trust me, there are a tons of them & growing). Even though I myself (and independent workers like me) stand to gain greatly from this, I recognize that not everyone will share my perspective because of the views from their view points. Let us try to analyze the perspectives in a relatively simpler way.

The benefits of ChatGPT, as everyone will vouch for, includes following.

1. Increased efficiency
2. Improved customer service
3. Enhanced decision-making
4. Cost savings
5. Increased innovation

With all these good things certain roles like CIO's, Chief Enterprise Architects and CISO's are skeptical about giving a free flow access to everyone within the organization. This goes to an extent where one may feel like labelling them not so forward looking or at an extreme, AI dumbs!

Well let us see their point of view; who knows, our perception may change!

• Unwittingly sharing confidential third-party or internal company information with ChatGPT or similar platform can lead to leakage of such sensitive data. This is because, with a continuously learning model, such data is perceived to be absorbing into data model of these platforms, which can then be spitted out to others with relevant queries.
• Privacy is a crucial concern when it comes to such platforms (OpenAI or others). While the tool does currently provide warnings to users about inputting sensitive or personal information, it remains unclear whether the creators of the tool are fully compliant with international privacy laws or have put in place adequate controls to safeguard user data and ensure respect for individual data rights(I believe they are but .. there are regulatory bodies who has to believe as well ?? ).This lack of clarity is a valid concern and highlights the need for greater transparency and accountability around data protection measures within the ChatGPT or similar platform.. Such unauthorized sharing of confidential information online including with ChatGPT, can potentially violate an organization's security policies and harm its reputation.
• The concerns of CIOs, Chief Architects and CISOs over ChatGPT mainly stem from the fact that the owners of this third-party chatbot are not within their organization's control. If a security breach occurs in ChatGPT, sensitive information that the company is bound to protect by law or contract can be exposed, resulting in reputation damage. This is a plausible scenario that must be taken into account when considering the risks and benefits of using ChatGPT. [ Just an Important note here - By virtue of microsoft's stake in OpenAI, now chatGPT is served from Azure [here is a link to Microsoft site that may be of interest], I believe. But OpenAI is still an Independent company]
• ChatGPT, being a third-party system and has the perceived ability to soak up data like a sponge. And even if the security of this AI chatbot remains intact, there's still a risk of company employees unintentionally spilling the beans on confidential customer or partner information. This possibly can land the company in hot water since they are contractually or legally obligated to keep such information under lock and key.
• The issue of content ownership in ChatGPT can get quite complicated. The terms of service state that the output generated by ChatGPT belongs to the person or service that provided the input. However, things can get dicey when that output includes information that is legally protected and was obtained from other sources no matter how much dated they may be. This can lead to all sorts of complications and potential legal headaches.
• Using ChatGPT to generate written material that possibly is inspired by copyrighted property, including licensed open source materials, raises concerns about potential copyright violations. For example, in the scenario of code [Just have a look at number of ChatGPT extensions in vsCode ] if ChatGPT is trained on open-source libraries and then outputs code when answering a question, a developer might use that code in software products that a company ships. This could put the company in terrible consequences. [Having said so I believe, CIOs, Chief Architects & CISOs must look at the current code that is being produced which are human written]

So look, the concerns of CIOs, Chief Architects and CISO's are not so invalid. They are neither hesitant nor ignorant - They just need time to weigh the risks and I believe they should take the time!

My Observations so far with ChatGPT -

1. It does not remember details given by user beyond the conversation.
2. Its memory is cut-off beyond September 2021 and can easily be verified.
3. It uses the conversational info to build intelligence not memory i.e. it analyzes sentence structure to answer questions based on past memory i.e. before sept 2021.
4. It looks to me that it makes reasoning errors occasionally. Probably because it is trained on a data set that includes conflicting data/information on same topic.
5. It may apologize, like a real human being, when you catch an error and tend to agree with what you are saying.
6. It denies answers to questions that may fall in ethical realm.

Having said so I would encourage you to read the information on this page of Open AI . It gives a lot of info on how this thing is progressing.

I believe trust is an important factor in this whole scenario. We hope that better understanding/agreement develops between platform providers such as OpenAI and consuming Enterprises who plan to use it. Bringing organizations with a greater degree of involvement in the development and implementation process of ChatGPT or equivalent platforms, could augment their comfort and confidence levels in utilizing the chatbot. Consequently, it can lead to a broader acceptance and implementation of such technologies in the workplace. I foresee partnerships getting established between IT Service Providing Organizations & OpenAI!

Moreover, from the perspective of platform providers like OpenAI, for their platform(s) to be widely used, they definitely need to address the concerns by getting their platform certified by data security and privacy regulatory bodies. We can not forget the initial concerns that we had when Cloud was getting into enterprises, and how it was resolved when cloud providers got them shelves certified to meet even the stringent privacy & data protection regulations.

Thirdly, I believe, enterprises may find themselves in a dire need to optimize or install more sophisticated content filtering systems at their network perimeter, given the substantial attention that OpenAI and other similar platform providers have garnered. As the outbound data volume requiring filtration is poised to skyrocket, organizations may have to invest extensively in a distinct deployment of advanced technology. Well, this definitely will take some time, after performing some deep cost-benefit analysis!

So shall not we be patient with our CIO,CEA and CISOs? Let us also adore the unexpressed fact that they are also concerned about the impact on people & their jobs!

Yes we should, but one thing must be considered- More agile companies win! So do small and mid size companies have an edge here?

What do you say?

Do contribute with your thoughts!

Cheers!