ChatGPT doesn’t cease to amaze and disrupt – this time it is about making random encounters in computer games more interesting and making their world seem more alive. When it comes to the real world – Natural Language Processing is tackling the building blocks of life itself.
And yet this is not enough for Tooploox CS and AI news. More is to be said about demand generation and big tech pushing the world forward.
ChatGPT brings more life to computer games
Mount and Blade is a game that allows players to be a part of the living world of Caldaria. The world resembles medieval-era Europe and the gameplay blends elements of horseback fight simulation, real-time strategy, and role-playing.
The world has its own politics, distinct factions and powers, and the player has the liberty to join any of them or stay independent, as a mercenary, warlord or bandit. Yet the game lacked role-playing elements, with characters having little or nothing to say.
This challenge has been tackled by one of the community’s modders, who fed ChatGPT knowledge about the game world and allowed all in-game characters to talk using it – with a player entering prompts from their keyboard, rather than choosing from options. This resulted in the players having free conversations with blacksmiths about the art of smithing and with the noble families on the intricacies of politics.
While the chat mechanic lacked the wit and personality of prewritten dialogues authored by professionals, it can be an interesting attempt to make a gaming world more immersive and deliver a better experience for players.
The demo version can be seen in the video below.
OpenAI and Microsoft extend partnership
OpenAI’s face-melting ChatGPT hasn’t stopped amazing companies and individuals around the world since early December 2022. This month brings another investment from Microsoft that aims to support the organization’s development of new models and technologies. Microsoft is also participating in building and delivering more computing power and resources required to build next, more sophisticated models.
Full information about the partnership can be found through the link below:
Report: 85% of demand-generation campaigns involve chatbots
According to a report issued by Botco.ai, up to 85% of all demand-generation campaigns involve chatbots of various kinds. Also, nearly all marketers (99% to be precise) who use chatbots in their demand generation campaigns claim that this technology increases lead-to-client rate. For 56% the increase was by 10% or higher.
More details can be found in this full report.
Natural Language Processing (NLP) models capable of designing non-natural proteins
Scientists from the University of California, San Francisco came up with a model capable of designing proteins that actually don’t exist in natural environments, effectively designing them from scratch. Some of these proteins performed similarly to natural ones, even when significantly diverting from their natural counterparts.
The technology used to design them is comparable to Natural Language Processing techniques, where the model aims to predict the next word in a phrase sequence. In this particular case it is about predicting the next amino-acid in a chain which is later used to sequence a protein.
The applications of these artificial proteins vary from designing new pharmaceutical products to degrading plastic – in fact this field is still brand new and the full possibilities are yet to be explored.
Brace yourselves! DetectGPT is here
ChatGPT was released in December 2022 and doesn’t cease to amaze users. On the other hand though, the tool can be used for multiple malicious purposes, including spreading fake news or plagiarism. The latter has become a serious problem for educational institutions, where students frequently use the tool to generate essays.
Flooding the research and educational world with plagiarized content generated by AI presents an existential threat. To tackle this problem, researchers from Stanford University have delivered an AI-based solution that detects ChatGPT’s shenanigans in texts. The solution will be available soon and the paper can be found here.