A NSFW AI that made available as opensource faces a lot of ethical, technical and security challenges. The source code of open-source software is made freely available to anyone who wants to view, modify it and distribute the changes. This level of visibility is great for the innovation and collaboration, but it starts to have risks when you are dealing with more sensitive technologies — like NSFW AI. Abuse fears Along with the press release, a number of industry experts cautioned that open-sourcing NSFW AI is more dangerous than deploying it behind-the-firewall as bad actors can take the code and tamper with its settings — potentially to bypass safeguards, resulting in obscene content slipping through undetected.
Open-source facial recognition software struggles illustrate what this danger might look like at a historic level. In 2020 IBM offered facial recognition system became notorious and was criticised as being abusive & invasive of privacy. The same vulnerabilities apply to NSFW AI as well, with open-source versions at risk of being abused by those looking to bypass moderation on platforms such as Twitter which processes 500 million sweats a day.
There are also some hefty economic ramifications to making NSFW AI open source. Building and maintaining these AI systems cost a lot of money, with companies like Facebook or Google pouring hundreds millions into research, development and data buying. Shari Steele, executive director of the Tor Project: “If these systems are released as open source., it would decrease ROI and possibly remove incentive for new innovations in content moderation technologies.
Efficiency and accuracy are highly important for a NSFW AI so it may be difficult to uphold these standards under an open-source model. Community contributions — based on open-source, I mean we have bugs thorough the systems. This means that even an open-source NSFW AI system would need a lot of governance and regular updates to keep it above 90% detection accuracy, the type of oversight you see best managed by teams who are heads-down working in privacy-sensitivplayces.
The AI luminary Elon Musk has warned specifically about open-source AI, saying that “Open-sourcing general artificial intelligence will have unintended consequences” which could mean looking for ways to dismantle it or manipulate the system human nature. This view emphasises the trade-off between transparency and security, in relation to open-source NSFW AI.
Nsfw ai isn't impossible to open source, but the interplay between innovation, security and ethics is complex. Although open-source software encourages collaboration and transparency (and is designed to discourage misuse, inaccuracies, or promote cost reduction), the drawbacks can be subverting. Whether or not nsfw ai should be open-sourced is something that needs to take into account the possible effects on content moderation and user safety.