In our blog post of April 24, 2023, linked to here, we discussed the Good, the Bad and the Ugly of ChatGPT (indeed the article received some ‘thumbs up’ comments from ChatGPT itself). In it, we also commented on the Samsung case, where for those users who input company or proprietary information into ChatGPT, they risked leaking sensitive information to a third party.

Further, we commented that if you are developing potentially patentable subject matter (ie inventions) there are novelty and inventive step requirements that must be met for a patent to be granted. Novelty is assessed against the prior art (all publicly available material) which can include the inventor’s own disclosures. By placing information into ChatGPT, a business risks creating a novelty-destroying disclosure that would have an impact on the validity of patent protection.

Since this article we have now had at least two enquiries from new inventors who have used ChatGPT to ‘hone’ their invention. As mentioned above, this may lead to a risk of disclosure, plus others. In this article we explain those risks, why and how they may occur, and offer some solutions which may help.

 

Read More Insights Here:

 

So How Does ChatGPT Work?

Like other large language models (LLMs), ChatGPT is what is known as a generative AI (gen AI) model that accesses vast quantities of real data from the internet to respond to user queries, known as ‘prompts’.  What makes ChatGPT unique among other gen-AI models is its super-fast and human-like response capabilities.  ChatGPT has such capabilities because it was pre-trained on a large collection of data, including conversational data, which allows the technology to understand the context and flow of a prose.

 

So, What’s The Problem?

  1. Novelty Destroying Disclosures

ChatGPT saves all data entered by users and all data produced by ChatGPT in response[1].  ChatGPT also collects user IP addresses, browser types and settings, and data on user interactions with the site—such as type of content engaged, features used, and actions taken.

In other words, anything you type into ChatGPT may be saved and possibly used when generating responses for other users. This therefore could be considered a novelty -destroying disclosure.

  1. Non-Human inventors

Further, many IP laws around the world require an inventor to be a human (a ‘natural person’) and in some countries, including the US, declarations must be made at the time a patent application is made that the inventors are all stated and that all inventors are natural persons. The USPTO have recently shared some guidance on this[2] and the DABUS[3] case is an example of where this is currently being tried and tested.

  1. Breach of confidential information

Many professionals are subject to confidentiality restrictions stemming from non-disclosure agreements (NDA) or confidentiality clauses in commercial agreements. If any of these restrictions apply to you, you should not enter any related confidential information into the platform, as entering that information into ChatGPT may breach your obligation to protect the confidentiality of that information.

Lawyers and legal professionals have a further, ethical, responsibilities to maintain attorney-client privilege and the confidentiality of client information and therefore may not enter any such information into ChatGPT.  Medical professionals should also avoid putting patient information into ChatGPT to avoid inadvertently disclosing patient medical information or committing a HIPAA[4] violation.

 

Solution 1 – Opt Out

While it is advisable not to enter confidential information into ChatGPT, the AI recently unrolled a new feature[5] that allows you to disable your ChatGPT history, so your conversations aren’t used to train its model.  To opt out, enter ChatGPT and:

  • Choose ‘Settings’ from the menu.

  • Click ‘Show’ next to ‘Data Controls’ to expand the window and then toggle the switch next to ‘Chat History & Training’ to the off position to stop your data from being used to train ChatGPT.

Remember, opting out will still not protect your confidential information, so be advised not to enter any sensitive information into ChatGPT, as OpenAI will still store your interactions with ChatGPT for 30 days and review those transcripts, even if you turned off your chat history.  For this reason, make sure to delete or ‘anonymize’ any sensitive information (change names, addresses, financial information, etc.) before putting it into ChatGPT.

 

Solution 2 – Use Proprietary Tools and LLMs

There are more and more ‘private’ LLM models and solutions being launched daily that can be used to do various jobs, from creating images, to prose, to inventions, to code. From cactus.ai to DALL-E to Mistral, they are at various stages of development, and some do require some level of IT skills to set up and manage on a private network within your own domain, so that your data is kept private and secure.

Indeed, OpenAI have also recently launched an Enterprise version of ChatGPT which claims to solve some of the problems outlined. However, one main drawback mentioned above still remains – that is the requirement to be a human inventor.

A long -established solution offered by a company in the market called IPRova[6] gets around this problem by using humans to review IProva output and to perform the invention. Basically, their software identifies ‘high-value’ areas for invention based on points of market and technology disruption in the area in which you would like to invent. This may be based on a business strategy, product roadmap, experience journey or other more general target area definitions. Selected target areas can then be populated with inventions through an invention creation campaign – and this latter part is entirely done by humans.

However, as there is a level of human involvement, there is a price tag for these inventions.

 

Solution 3 – Do It the Old-Fashioned Way

There’s nothing like blood sweat and tears in defining an invention, in fact I would argue that defining the invention is the teensiest tiniest part of the problem, as the greatest ‘value’ is in defining the product market fit, building an MVP (minimum viable product), building  a great team around you, obtaining funding and managing cash flow, bringing the product to market, redefining the product based on customer feedback, and making the whole end to end product lifecycle sustainable and repeatable.

Whether or not you create the invention yourself from scratch, you still need to protect it and commercialise it and as yet, you can’t shortcut those steps.

Further, using ChatGPT has already been shown to potentially diminish creativity and critical thinking skills[7], so you may better off overall using pen and paper to start with.

 

[1] https://help.openai.com/en/articles/5722486-how-your-data-is-used-to-improve-model-performance

 

[2] US example

[3] https://en.wikipedia.org/wiki/DABUS

[4] https://www.hipaajournal.com/hipaa-compliance-checklist/

[5] https://help.openai.com/en/articles/7730893-data-controls-faq

[6] https://www.iprova.com/

[7] https://time.com/7295195/ai-chatgpt-google-learning-school/

 

WHERE INNOVATION THRIVES
We put you in control of your destiny by organising, optimising, & monetising your Intellectual Property.
chatgpt enterprise IPStrategy