«Should we use ChatGPT?» we asked ChatGPT.
- Intellectual Property Rights: As an encyclopedia, the content provided by Chat GPT needs to be carefully reviewed to ensure it doesn't infringe on any copyright or intellectual property rights of others. Using copyrighted material without proper authorization can lead to legal issues.
- Accuracy and Liability: As an encyclopedia or search function, the information provided by Chat GPT should be accurate and reliable. If users rely on the information and suffer harm or damages as a result, there could be potential liability issues.
- Ethical Considerations: When integrating AI models like Chat GPT, ethical considerations should be taken into account, including preventing the dissemination of harmful or discriminatory content...»
(Source: ChatGPT. ChatGPT’s source? Not sure...)
Our considerations and our thoughts on ChatGPT’s reply:
Intellectual Property Rights
We believe that the legal aspect, especially with regard to copyright - should not be underestimated: just as an agency relying on the fact that our work is protected by copyrights, this should also apply to the work of others that is cited by chatbots.
Accuracy and liability
This answer implies that the information provided by ChatGPT may not be accurate or even true. In fact, it may even be fabricated. Have you heard about the legal case in a Manhattan court, Mata (an individual) v. Avianca (an airline)? In this case, the lawyer cited more than half a dozen relevant court decisions. There was just one problem: No one - neither the airline's lawyers nor the judge himself - could find any of the decisions cited in the brief. That was because ChatGPT had made them all up (source: New York Times). While this says a lot about ChatGPT's ability to be creative, it is also telling in terms of factual accuracy, or rather inaccuracy.
We think these concerns are absolutely justified. Although legal problems in this regard could probably be prevented in our jurisdiction with a disclaimer, there is only very limited control over what sort of content is generated and what happens with it. And if it’s not a question of legal liability then it is definitely an ethical one.
While AI isn’t really all that new, with the sudden public attention that OpenAI has attracted with the release of ChatGPT, Microsofts investment, Google’s launch of Bard and Apple’s rumored activities in the field (source: Bloomberg), politicians and regulators around the world have become aware of potential risks and problems. While it took 14 years since the beginning of Facebook for its founder to appear before a US Senate committee, it took less than 6 months after the public launch of ChatGPT, for OpenAI founder Sam Altman to testify before the US legislators.
Of course, it also has to do with the rapid spread of the technology, but it is nevertheless an indication that a lot is in motion and a lot will still change - on the development side as well as in the legal area. At Reach, we will continue to follow the developments, and in the meantime learn and use this technology that opens up so many possibilities.
And for now: No, we do not recommend integrating a ChatGPT-based encyclopaedia into a website.