Webdesk: Meta Platforms Inc. said Monday it will continue to offer its artificial intelligence tools to authorised academics despite internet discussion board reports that its latest large language model had leaked to illegal users.
The technology giant stated that the current release strategy balances accountability and transparency.
Meta, owned by Facebook, unveiled LLaMA, an AI research project, last month. Meta stated that the model can mimic OpenAI’s ChatGPT and Alphabet Inc.’s human-like conversational abilities with less computational power.
Meta’s AI research unit publishes most of its work, unlike OpenAI, which charges software developers for access. There are possibilities of spreading misinformation using AI tools.
Meta grants non-commercial licences to researchers and other government, civil society, and academic groups following vetting to prevent misuse.
The model was allegedly uploaded on 4Chan last week. Reuters could not confirm them.
According to Meta, They were handling its LLaMA release similarly to previous models. Meta said that it will not change its strategy.
“Meta wants to share state-of-the-art AI models with the research community to help us evaluate and enhance such models,” Meta added.
