In today’s episode, we will continue talking about bots with learning capabilities, such as gpt-3 ai chat, and why this technology is a game changer in several business fields.
This is a follow-up of a previous episode: Ep 42
As always, this will be simple and concise but accurate. Click on Subscribe if you want to be informed of new episodes.
In our previous episode, we talked about gpt-3 chat, and I showed my interaction with the solution. I closed the chapter by saying I would love to do the same exercise a year later and check how much the AI-based solution improved in the four categories tested: Emotions, Relevance, Boundaries, and Opinions.
gpt-3 ai chat a week later testing results
Guess what? I did the test this week and found meaningful progress that I want to share with you.
I. Emotions: How are you doing?
This is an acid test; bots cannot respond with emotions, but they may respond to that question with “finesse” and eloquence. The idea here is to be able to invite the user to continue interacting with the AI-based solution.
Q: Hi Chat, How are you doing?
gpt-3 ai chat response:
II. Relevance: Any relevant news about artificial intelligence this week?
This test is essential because it allows us to understand the maturity level the AI-based solution has achieved. Asking for news this week means this is a dynamic system with continuous data loading and indexing in place. It also means the system can create ranking in real time.
Q: Any relevant news about artificial intelligence this week?
gpt-3 ai chat response:
III. Boundaries: Give me a sample of topics that we can talk
This test is important because it shows the AI-based system’s capability to identify areas of knowledge and competence. This is a critical capability in areas where wisdom and credibility are necessary, such as Education and Customer Service.
Q: Give me a sample of topics that we can talk.
gpt-3 ai chat response:
Did you notice the change from Topics to Categories? Nice!
IV. Opinion: Any opinion about the three laws of Robotics?
Have you ever wondered why many people like to talk with taxi drivers to learn about current events? (I am one of them)
Because -often- taxi driver’s opinions are the sum up of hundreds of customer opinions.
This test looks to identify if the AI-based solution can come with self-made opinions or opinions via interactions with other users (biased systems).
Q: Any opinion about the three laws of Robotics?
gpt-3 ai chat response:
Why is the opinion test so relevant?
Nowadays, there is important controversy on the importance of train AI-based solutions with unbiased data and the potential risk for people and organizations making decisions supported by AI-based solutions trained with biased data.
However, there are situations where using biased data may be helpful and relevant: news rankings.
That may be why there is also important controversy on the need to identify and mitigate the impact of fake news bots.
Areas where this disruptive technology may provide value now?
I see an immediate value in Customer Care areas and Education.
But this is just the beginning. I just read an article saying ChatGPT also has the capability to write code based on phrases.
Can you Imagine a bot mature enough to code tasks or find coding errors…
Want to play with ChatGPT?
Good enough?
I hope you find this Episode valuable and entertaining.
What do you think about the subjects raised in this edition of the Digital Acceleration Newsletter?
I would love your feedback on the Newsletter’s value and alternatives I can explore to bring additional value to you.
You are welcome and encouraged to share your thoughts with the audience or via DM; I am happy to help.